Analyzing Students Attention in Class Using Wearable Devices

Size: px
Start display at page:

Download "Analyzing Students Attention in Class Using Wearable Devices"

Transcription

1 Aalyzig Studets Attetio i Class Usig Wearable Devices Xi Zhag 1, Cheg-Wei Wu 1, Philippe Fourier-Viger 2, La-Da Va 1, Yu-Chee Tseg 1,* 1 Departmet of Computer Sciece, Natioal Chiao Tug Uiversity, Hsichu, Taiwa 2 School of Natural Scieces ad Humaities, Harbi Istitute of Techology, Shezhe, Chia zhag@cs.ctu.edu.tw, cww3@ctu.edu.tw, {ldva, yctseg * }@cs.ctu.edu.tw, philfv@hit.edu.c Abstract Detectig studets attetio i class provides key iformatio to teachers to capture ad retai studets attetio. Traditioally, such iformatio is collected maually by huma observers. Wearable devices, which have received a lot of attetio recetly, are rarely discussed i this field. I view of this, we propose a multimodal system which itegrates a headmotio module, a pe-motio module, ad a visual-focus module to accurately aalyze studets attetio levels i class. These modules collect iformatio via cameras, accelerometers, ad gyroscopes itegrated i wearable devices to recogize studets behaviors. From these behaviors, attetio levels are iferred for various time periods usig a rule-based approach ad a datadrive approach. The former ifers a studet s attetio states usig user-defied rules, while the latter relies o hidde relatioships i the data. Extesive experimetal results show that the proposed system has excellet performace ad high accuracy. To the best of our kowledge, this is the first study o attetio level iferece i class usig wearable devices. The outcome of this research has the potetial of greatly icreasig teachig ad learig efficiecy i class. Keywords Activity Recogitio; Attetio Sesig; Body-Area Network; Machie Learig; Wearable Computig I. INTRODUCTION The Iteret of Thigs (IoT) [2] ad wearable techologies [16] have rapidly become key research areas i computer scieces, as they have multiple real-life applicatios i a wide rage of domais. A wearable device is a microelectroic computig ad sesig system that ca be comfortably wor by its users. Its embedded software ad sesors allow measurig users vital sigs ad ambiet coditios. I cotrast to traditioal computers, wearable devices are lightweight, small, geerally iexpesive, ad close to the wearer s bodies. They are thus highly promisig for collectig ad aalyzig huma behavioral data. I the techology idustry, may of the largest compaies such as Google, Apple, ad Microsoft have recetly desiged wearable devices. Numerous traditioal accessories have bee trasformed ito wearable devices [16], icludig glasses, watches, clothes ad eve diapers. As poited out i a report by the Iteratioal Data Corporatio [1], the worldwide shipmet of wearable devices is expected to reach millio uits by 19. Numerous research studies have bee devoted to wearable applicatios, such as detectig users activities [5], social actios [8], falls [24], ad users browsig behaviors i retails [21]. Although wearable techology has bee applied i several domais, it is rarely cosidered for measurig studets attetio levels i class. Attetio ca be defied as the behavioral ad cogitive process of selectively cocetratig o certai pieces of iformatio [1]. I a previous study [23], it was foud that selective ad sustaied attetio has a sigificat impact o learig. Learers attetio is closely related to their learig efficiecy ad learig outcomes. Kowig the attetio levels of studets i class would thus greatly beefit both studets ad teachers. This iformatio ca help studets better uderstad their learig processes ad adapt their learig strategy, while teachers ca use this feedback to gauge studets iterest ad adjust teachig strategies to capture ad maitai studets attetio. Traditioal ways of measurig studets attetio levels iclude fillig questioaires [22], performig experimetal tests [13], ad direct observatio [14]. Although these methods are suitable i some scearios, they are time-cosumig, require huma itervetio, ad most of them caot be applied i real-time, or to a large group of studets. Some studies [3] have used surveillace cameras ad face recogitio models to automatically measure studets attetio levels. However, the accuracy of these approaches is substatially iflueced by factors such as lightig, camera positios, ad backgroud iterferece. Some studies [3] used eye trackers to detect visual attetio. However, these devices are quite expesive ad prologed use may cause eye ijury [19]. To address these drawbacks of curret approaches, this study desigs a approach for measurig studets attetio levels i class usig low-cost wearable devices. Achievig this goal is however challegig sice attetio is a iteral cogitive process that ca oly be idirectly observed through its effects o exteral actios, such as motio chages, visual focus, ad physical behaviors. To tackle this challege, this paper proposes a ovel system that captures head-motio, pemotio, ad visual-focus data for attetio iferece. The major challeges of this research are listed below. 1. Recogizig users differet activities requires the use of multiple sesors, which geerate heterogeeous data. As more types of sesors are used, it becomes icreasigly difficult to joitly aalyze the heterogeeous data that they geerate to build a iferece model 2. Idetifyig the features to be used to build a iferece model is also challegig. Huma activities are usually complex ad ca be described usig umerous features. Whe both physical ad psychological data are collected, feature selectio becomes eve more difficult.

2 3. Collectig accurate ad reliable data is ofte difficult i real-world applicatios. Data ca be oisy. Moreover, i may situatios, data are streamig at various speeds ad are ubouded. Also, real-time applicatios require fast processig time. 4. Aother challege is that persoal differeces may ifluece the success of attetio detectio, as differet persos may behave differetly i the same situatio. Moreover, some participats i the experimetal evaluatio may lie by pretedig to have bee focused whe they were distracted. To address the above issues, we propose a ew system that itegrates a Head-Motio module, a Pe-Motio module, a Visual-Focus module, ad a Active-App module for attetio iferece. These modules ru o iexpesive hardware, ad rely o machie learig techiques to select appropriate features to recogize attetio-related activities, such as head motios, had motios, ad visual behaviors. These modules are implemeted o the Raspberry Pi [] platform. I particular, the J48 decisio tree learig algorithm [27][28] is used to trai the motio classificatio models usig discrimiative features of sesor data. The behaviors recogized by these models are the fed to a attetio iferece egie, which cosists of two differet attetio iferece algorithms. Oe applies a rule-based approach, ad the other employs a data-drive approach. Fially, visual reports are geerated ad preseted to the users. Extesive experimetal results show that the proposed system is efficiet ad accurate. The outcome of this research has the potetial of greatly icreasig teachig ad learig efficiecy i class. The remaider of this paper is orgaized as follows. Related work is reviewed i Sectio II. Sectio III presets the proposed system. Sectio IV evaluates the performace of the proposed system. Fially, Sectio V draws the coclusio ad discusses future work. II. RELATED WORK This sectio first reviews the mai studies o attetio detectio. The, it takes a broader perspective to review other models of huma activity recogitio. Traditioal methods for measurig attetio levels ca be geerally categorized ito three types. The first type cosists of askig learers to fill questioaires [22]. The secod type is physiological experimets [13], where experts observe the reactios of learers whe performig tasks, to ifer their attetio levels. The third type is direct observatio [14], where experts evaluate the thought process of learers based o data recorded durig a set period of time usig devices like video ad voice recorders. Although traditioal methods are commoly used, a major drawback is that they require huma itervetio. Various devices have bee used for attetio iferece: 1. Electroecephalography devices [15][17] record the electrical activity of huma brais. Although they ca measure brai waves of a perso, they igore other aspects such as body movemets ad visual focus. 2. Eye trackers [29][3] detect eye movemet trajectories. However, they ca be quite expesive ad log term use may cause eye ijuries [19]. 3. Video cameras [3] have bee used to detect facial expressios ad body movemets. However, their use raises privacy issues, requires more computig power, ad their accuracy is iflueced by factors such as lightig coditios, camera positios, ad backgroud iterferece. I terms of huma activity recogitio, solutios ca be categorized ito sesor-based ad visio-based approaches. Sesor-based approaches rely o body-wor iertial sesors to ifer physical activities ad lifestyles [11]. Daily routies of wearers have bee detected usig topic models [9]. I aother study, sigle body-wor accelerometers were utilized to recogize social actios, icludig speakig, laughig, gesturig, drikig, ad steppig [8]. Body-wor accelerometers have also bee cosidered to recogize household activities for cotext-aware computig [5]. A lot of visio-based solutios have bee proposed. A camera-based surveillace system was desiged for detectig huma movemets [18]. Other researchers [3] have explored the relatio betwee head rotatio, eye gaze directio ad facial features to ifer huma attetio. To reduce the computatioal cost for image processig, a collaborative model usig a depth camera ad a iertial measuremet sesor was proposed [6]. The model uses esemble classifiers at both the feature levels ad decisio levels. III. ATTENTION INFERENCE SYSTEM A. Classroom Sceario This study cosiders the followig sceario (Fig. 1): (1) A teacher shows slides for teachig usig a projector. Each slide cotais a special mark, such as a school logo, used by our system to check if each studet is payig attetio to the slides. (2) The teacher has a computer i frot of him/her, which acts as a server. It collects sesor data ad ifers studets attetio levels. (3) Each participatig studet sits at a desk ad wears smart glasses or a smart cap. The wearable device is i either case equipped with a camera, a accelerometer ad a gyroscope. (4) Each studet uses a smart pe to take otes as usual. A SesorTag [26] embedded with a accelerometer ad a gyroscope is attached to each pe. (5) Alteratively, a studet may also take otes o his/her otebook. It is the assumed that otes are iserted i a PowerPoit file or usig a PDF editor. (1) (2) (1) Projectio image (with logo) (2) Server & iferece egie (3) Smart glasses/cap (4) Smart pe with SesorTag (5) tebook (or Tablet) (5) (4) (3) Fig. 1. The classroom sceario cosidered i this work. (1) projectio image ( (2) server & iferece (3) smart glasses/cap (4) smart pe (5) otebook (PPT or

3 Server Side Cliet Side Attetio Level Iput Process Output Logo of slides, provided by the teacher i advace Image data captured by the smart cap s camera ACC ad GYRO data geerated by the smart cap s MPU 5 ACC ad GYRO data geerated by the smart pe s SesorTag App logs from otebook ad tablet Head-Motio Module Visual-Focus Module Pe-Motio Module Active-App Module Attetio Iferece Egie Attetio Level Report Time Activity Report Head motio = {Still} Head motio = {Dow} Head motio = {Up} Time Fig. 2. Hardware compoets of our prototype system. B. Prototype Desig This sectio presets the desig of our prototype system, based o commodity hardware ad a cliet-server architecture. O the cliet side, a wearable device implemeted o a Raspberry Pi Model B+ [] has bee desiged for sesig motio ad the visual focus of users. It itegrates a camera ad a sesor board MPU 5 cotaiig a 3-axis accelerometer ad a 3-axis gyroscope. The accelerometer ad gyroscope of MPU 5 are used to recogize head motios (e.g. raisig/lowerig the head), while the camera is used to capture the focused field of the wearer for visual-focus detectio. These compoets ca be easily itegrated ito a cap or glasses (see Fig. 2). A SesorTag is attached to each smart pe, ad cotais a 3-axis accelerometer ad a 3-axis gyroscope. The SesorTag o the smart pe is used to detect had motios (e.g. writig/still) of the user. The commuicatio betwee SesorTag ad Raspberry Pi relies o Bluetooth Low Eergy (BLE) protocol. O the server side, Wi-Fi protocols ad Apache Http Server Versio 2.4 [25] are used to collect data set from cliet side. The, the pe-motio, head-motio ad visual-focus modules are ru o server to recogize behaviors of users. The recogized behaviors are the fed to a attetio iferece egie to calculate users attetio levels. Fially, visualized reports are geerated by the egie to idicate each user s attetio levels for differet time periods. C. Overview of the System Architecture Fig. 3 shows the Iput-Processig-Output (IPO) model of the proposed system for a sigle user. It hadles streamig data usig a batch model. Durig each iteratio, it receives a batch of data, which icludes: (1) The image of the special logo o slides. (2) The image data captured by the user s smart glasses/cap. (3) 3-axis accelerometer (abbr. ACC) ad 3-axis gyroscope (abbr. GYRO) data geerated from MPU 5 of smart cap. (4) 3-axis ACC ad 3-axis GYRO data geerated from SesorTag of smart pe. (5) The logs of Apps used by the users. The system outputs a attetio level report ad a activity report, which idicate the attetio level of the user ad his/her recogized behavior durig each time period, respectively. The processig part of the system cosists of five modules, amely the Head-Motio, Pe-Motio, Visual- Focus, Active-App ad Attetio Iferece Egie modules. These modules commuicate based o a cliet-server architecture, as depicted i Fig. 4. These modules are explaied i the followig subsectios. Fig. 3. The IPO model of the proposed system for a sigle user. Pe-Motio Module Pe-motio Data Preprocessig Accelerometer Data Data Segmetatio Gyroscope Data Pe-motio Type Recogitio Pe-motio Feature Extractio Pe-motio Classifier Writig Still Other Recogized Behaviors Fig. 4. The architecture of the proposed attetio iferece system. D. Head-Motio Module Head motio is a exteral behavior that may serve as a importat idicator of studet attetio. Durig the data preprocessig stage, raw sesor data are sampled ad segmeted. There are six sesor sigal sources (received from a 3-axis ACC ad a 3-axis GYRO). The samplig rate is 5Hz. Data are partitioed ito uiform o-overlappig three-secod segmets. Segmeted data are set to the server. Let S = s 1, s 2,..., s be oe of the sesor sigal sources, where s i (1 i ) is the i-th sample i S. The proposed system extracts the followig six time-domai features related to eergy or magitude from S. Mea is a measure of cetral tedecy, which is defied as Mea S = s = i=1 s i. (1) Variace represets the expected squared deviatio from the mea, which is defied as Var(S) = Head-Motio Module Head-motio Data Preprocessig Accelerometer Data Data Segmetatio Head-motio Type Recogitio i=1 (s i s ) 2 Gyroscope Data Head-motio Feature Extractio Pe-motio Classifier Up Still Dow Attetio Iferece Egie Attetio Iferece Algorithms Attetio Scorig Fuctios Active-App Module Visual-Focus Module Attetio Level Report ad Activity Report. (2)

4 Root Mea Square (RMS) is the square root of the arithmetic mea of the squares of the values i S, which is defied as RMS(S) = i=1 s i 2. (3) Average Absolute Differece (AAD) measures the statistical dispersio of cosecutive samples i S, which is defied as AAD(S) = i=2 s i s i 1 1. (4) Zero Crossig Rate (ZCR) measures the umber of sig chages i S, which is defied as ZCR(S) = i=2 sig s i sig (s i 1 ) 2, (5) where sig(x) is a fuctio which returs -1/+1 if the sig of x is egative/positive ( is regarded as positive). Mea Crossig Rate (MCR) measures the umber of sig chages usig the mea as baselie, which is defied as MCR(S) = i=2 sig s i s sig (s i 1 s ) 2. (6) Sice there are two types of sesors (i.e., ACC ad GYRO) ad each sesor has three axes, calculatig the above features results i (2 3 6)=36 features. Moreover, for each type of sesor, the system also measures the covariace ad correlatio betwee each pair of its two sesor sigal sources. Let A = a 1, a 2,..., a ad B = b 1, b 2,..., b be two sesor sigal sources i a segmet. The covariace ad correlatio betwee A ad B are measured usig the followig two formalizatios. Covariace measures how much A ad B chage together, which is defied as Cov(A, B) = i=1 a i a (b i b). (7) Correlatio measures the stregth ad the directio of the relatioship betwee A ad B, which is defied as Corr(A, B) = cov (A,B) σ A σ B, (8) where σ A ad σ B are the stadard deviatios of A ad B, respectively. Sice there are two types of sesors ad each sesor has three axes, measurig the covariace ad correlatio betwee each pair of axes results i 12 features. Therefore, the system totally extracts (36+12) = 48 features. The, the 48 defied features are take as iput by the head-motio module to trai a classifier offlie, which is the used for olie recogitio of head motios. Three head motios are cosidered: {Still, Up (raisig the head), Dow (lowerig the head)}. The J48 decisio tree learig algorithm offered i Weka [27] is used for traiig phase. The traiig phase is performed i three steps. Step 1: Head-motio traiig data are collected ad each data is labeled with its class. Let C be the set of classes {Still, Up, Dow}. The result is a traiig set TraiDS = {(d 1, g 1 ), (d 2, g 2 ),..., (d m, g m )}, where d i (1 i m) is the i-th data istace ad g i C (1 i m) is the label of d i. Step 2: Each traiig istace d i is trasformed ito a feature vector f i described by the 48 defied features. This results i a trasformed dataset FDS = {(f 1, g 1 ), (f 2, g 2 ),..., (f m, g m )}. Step 3: This dataset FDS is fed to the J48 decisio tree learig algorithm to trai the head-motio classifier, called HM-Classifier(.). The traied classifier is the used for olie recogitio. As the server cotiuously receives segmeted head-motio data from cliets, it applies the HM-Classifier(.) to recogize the head-motio type of each segmet. This is performed i two steps. Step 1: For each received segmet, the 48 features are extracted, deoted as f SD. Step 2: HM-Classifier(.) is the applied to classify f SD. The output is its class label, deoted as HM-Classifier(f SD ). te that the umber of features ca be reduced by usig the gai ratio goodess fuctio [27][28] of the J48 classifier while preservig a high accuracy. E. Pe-Motio Module Pe motio is aother exteral behavior that ca help measure studet s attetio. The pe-motio module is desiged to detect if a studet is writig o a piece of paper. The SesorTag o the smart pe seds raw 3-axis ACC ad 3- axis GYRO data at a samplig rate of 5Hz to the smart cap. The smart cap collects ad segmets the received data ad the seds segmeted data to the server every three secods. The server the extracts the 48 features defied i Sectio III.D from the sesor data ad classifies the studet s pemotio data usig a classifier. This classifier is traied oce offlie. Three types of pe motios are cosidered: {Still, Writig, Other}. For the traiig phase, segmets of raw sesor data are maually labeled with these three classes. The traiig phase is doe usig the J48 classifier from Weka. The traied classifier is applied to olie recogitio every three secods to detect studet s pe motios. As previously explaied, the umber of classifier s features ca be reduced usig the gai ratio goodess fuctio [28]. By applyig this techique, the decisio tree learig algorithm selects oly the most discrimiative features for motio recogitio. I our implemetatio, this techique cosiderably reduces executio time for recogitio, while preservig a high accuracy. F. Visual-Focus Module The visual focus of a huma is also a importat idicator of the attetio, as humas ted to pay attetio to objects appearig i the ceter of their visual field. I our implemetatio, a camera is attached to the smart cap. The assumptio is that the camera captures what the studet sees, ad the images ca thus reveal what is curretly drawig the studet s attetio. I our desig, the camera is programmed to record image data oly whe the user s head is i motio.

5 Visual-Focus Module Head-motio Type Head-motio Type ={Up} Slide Detectio Had Detectio Head-motio Type = {Dow} Fig. 8. A example of ski color mask for had recogitio. Slide Foud Slide Had Foud Had Fig. 5. The system architecture of the visual-focus module. Fig. 6. Examples of slide detectio usig the SURF algorithm. Fig. 7. A example of feature poit detectio usig the SURF algorithm. Cosiderig the size of image data ad trasmissio rate, the Visual-Focus module is ru o the Raspberry Pi o the cliet side s smart cap. Fig. 5 shows the workflow of this module, which cosists of two sub-modules: Slide Detectio ad Had Detectio. (a) Slide Detectio If the head-motio type recogized by the Head-Motio module is {Up} (i.e., raisig the head), the Slide Detectio module will be triggered to idetify whether the user is payig attetio to the slides. Our system assumes that a school logo is show o each slide. If the logo is foud i a image captured by the camera, it is very likely that the slide is withi the studet s field of visio, ad that the studet is payig attetio to its cotet. To fid the logo from the captured image, we use the Speed-Up Robust Feature (SURF) algorithm [4] offered i the OpeCV library. Fig. 6 shows two examples of recogitio results usig SURF. As depicted i Fig. 7, SURF extracts feature poits from the logo image ad tries to match these poits with the captured image. If the umber of matched poits is greater tha a miimum matchig umber threshold, the system assumes that the slide is withi the studet s field of visio. Moreover, the system also verifies whether the slide is cetered i the studet s field of visio. If the slide is off-ceter by a distace greater tha a miimum off-ceter threshold, the system assumes that the studet is ot lookig at the slide. (b) Had Detectio If the head-motio type recogized by the Head-Motio module is {Dow} (i.e., the head is lowered), the Had Detectio module will be triggered to check if the user s had appears i the captured image. This iformatio will be further itegrated with the recogitio results of the pe-motio module to determie if the user is takig otes. The assumptio is that a studet s had should be visible i the images captured by the camera if he/she is takig otes, ad that it should ot be visible if the studet is lookig at the slides. To check if a had appears i a captured image, we apply a method based o ski color masks usig the YCbCr color space [7]. This method extracts all the potetially skicolored pixels from a captured image. If the Cb or Cr values of a pixel are i the [98, 142] or [133, 177] itervals, respectively, the pixel is cosidered as ski-colored [12]. For example, cosider the left part of Fig. 8, which shows a image captured by the camera. The right part of Fig. 8 illustrates the detected ski-colored pixels (represeted as white pixels). If the umber of ski-colored pixels is o less tha a miimum pixel threshold, the system assumes that the had of the user is visible i the captured image. G. Active-App Module Cosiderig that more ad more studets take otes o their otebooks ad tablets durig lectures, we foud that idetifyig active Apps that are curretly beig used by the studets is a efficiet way to ifer their attetio level. I this work, it is assumed that otes are take o PowerPoit, Word or PDF files. The Active-App module logs the Apps ad files curretly used o a studet s otebook or tablet, ad seds this iformatio to the server. The server uses this data to check if a studet is payig attetio to the course s slides. H. Attetio Iferece Egie Whe user s behaviors are detected (i.e., head motio, pemotio, visual focus, ad Apps used by user), the proposed attetio iferece egie will ifer the user s attetio level based o these behaviors. Two attetio iferece algorithms are proposed, called the rule-based approach ad the datadrive approach, respectively. (a) Rule-based Approach The mai characteristic of the rule-based approach is that it ifers studet s attetio levels based o user-defied decisio rules. These rules have the merit of beig ituitive, iterpretable ad to allow fast recogitio. Fig. 9 shows the workflow of the rule-based approach. The algorithm is a iterative method. I each iteratio, it takes the recogitio results of modules as iputs, ad outputs a attetio state. Two types of attetio state are cosidered: {Focused, Ufocused}.

6 Attetio State Focused Ufocused Start Is the user s head still? Still duratio > threshold? Recogized Behaviors High-level Feature Extractio Classifier Focused Ufocused Up What is the user s head motio type? Dow Fig. 1. Workflow of the data-drive approach. Does slide appear i the user s visual field? Does a had appear i the user s visual field? Head Motios Pe Motios Up Up Still Still Still Still Dow Dow Dow Dow Other Other Still Still Still Still Write Other Write Other Take otes by smart pe? Use class-related files o otebook or tablet? Foud Slide Foud Visual Focus Had Foud Had Slide Had Had Active Apps PDF PPT Fig. 11. A example of the iput of the data-drive approach. Focused Ufocused Focused Focused Ufocused Ufocused Fig. 9. Workflow of the rule-based approach. I a iteratio, the algorithm proceeds as follows. At the begiig, the algorithm checks whether the user s headmotio type is {Still} or ot. If it is true ad the still duratio is o less tha a threshold, the user is likely i a daze ad thus the algorithm outputs {Ufocused}. If the head-motio type is ot {Still}, the algorithm the checks whether the user s headmotio type is {Up} or {Dow}. If the head-motio type is {Up}, the algorithm uses the recogitio result of the Slide Detectio module to idetify whether the slide appears i the user s visual field. If the slide appears i the user s visual field, the algorithm outputs {Focused}. O the cotrary, it outputs {Ufocused} if the slide does ot appear i the user s visual field. If the head-motio type is {Dow}, the algorithm uses the recogitio result of had detectio to idetify whether the user s had appears i the user s visual field. If the had appears i the user s visual field, the algorithm uses the recogitio result of the pemotio module to idetify whether the user had writte or ot. If the recogitio result of the pe-motio module is {Writig}, the algorithm outputs {Focused}. If the recogitio result is ot {Writig} or the had does ot appear i the user s visual field, the algorithm uses the result of the Active-App module to check whether the user had used class-related Apps. If the result is true, the algorithm outputs {Focused}. Otherwise, {Ufocused} is outputted. (b) Data-drive Approach A o-empirical approach is to use a data-drive solutio. It uses machie learig methods to fid hidde iformatio i the collected data, ad utilizes this iformatio for buildig a decisio model for attetio iferece. If there are more useful traiig data beig used for model buildig, the iferece results will be more accurate. Fig. 1 shows the workflow of the data-drive approach. The algorithm is a iterative method. I each iteratio (also called widow), it takes -secod recogitio results of the Head-Motio, Pe-Motio, Visual- Focus, ad Active-App modules as iputs. I our implemetatio, is set to 3. The, the algorithm extracts high-level features from the iput data ad classifies the studet s attetio state (i.e., Focused or Ufocused) usig a classifier. The classifier is traied offlie. For the traiig phase, the class of each traiig istace is maually labeled. The traiig phase is doe usig the J48 classifier from Weka. The traied classifier is applied to olie recogitio every secods to detect studet s attetio state. Next, we itroduce the high-level features extracted by the data-drive approach. Let H 1, H 2, H 3, ad H 4 deote the - secod recogitio results of the Head-Motio, Pe-Motio, Visual-Focus, ad Active-App modules, respectively. Let H i = h i1, h i2,..., h i, 1 i 4, where h ij (1 j ) is the j-th data istace i H i. Data istaces i H i are also called activities. For example, i Fig. 11, the 3-secod recogitio results from the head-motio module is H 1 = Still, Still, Dow, Up, Dow, Up, Still, Still, Dow, Dow. Let R i = {r i1, r i2,..., r Ri } be the set of activities i H i. For example, H 1 has three types of activities {Still, Up, Dow}. The, for each iput H i (1 i 4), the system extracts the followig high-level features. The umber of two adjacet data istaces i H i that are differet activities. For example, i the aforemetioed H 1, that umber is 6. The umber of activities that are r ik (1 k R i ) i H i. For example, the umber of activities that are {Dow} i H 1 is 4. The maximum duratio of r ik (1 k R i ) i H i. For example, the maximum duratio of {Dow} i H 1 is max{1, 1, 2} = 2. The miimum duratio of r ik (1 k R i ) i H i. For example, the miimum duratio of {Dow} i H 1 is mi{1, 1, 2} = 1. The average duratio of r ik (1 k R i ) i H i. For example, the average duratio of {Dow} i H 1 is (1+1+2)/3 = (c) Attetio Scorig Fuctio The attetio scorig fuctio is used to idicate the stregth of the studet s attetio level i a visual maer. Recall that i our proposed attetio iferece algorithms, i each iteratio or widow, they will output a attetio state (i.e., {Focused} or {Ufocused}). Therefore, i the -th iteratio, attetio states are obtaied. Let ST = st 1, st 2,..., st be the ordered set of these attetio states. The attetio score of ST is defied as Score(ST) = 1 i=1 g(st i ), (9) where g(s i ) returs a value g 1 (g 1 ca be defied as a positive value) if st i is {Focused}, ad returs g 2 (g 2 ca be defied as a egative value or zero) if st i is {Ufocused}. I additio to the above attetio scorig fuctio, teachers ca desig other iterestig scorig fuctios depedig o the requiremets of the applicatios.

7 Accuracy (%) Accuracy (%) Accuracy (%) Accuracy (%) IV. EXPERIMENTS AND PERFORMANCE EVALUATION Experimets were coducted to assess the performace of the proposed Head-Motio module, Pe-Motio module, Slide Detectio module, ad the two attetio iferece algorithms. A. Performace of Head-Motio Module To assess the Head-Motio module s ability at recogizig head motios, a experimetal study with five participats was performed i a classroom eviromet. Three head motios were cosidered: {Still, Up (raisig the head), Dow (lowerig the head)}. To simulate a real classroom eviromet, participats were asked to sit at a desk ad wear the desiged head-mouted wearable device, which cotais a 3-axis accelerometer ad a 3-axis gyroscope. A video camera was used to record the sessio for groud truth labelig. Totally, 25 data istaces were collected. Amog those, 15 data istaces were used for the traiig phase ad the other for the testig phase. Characteristics of the traiig ad testig datasets are preseted i Table I. I Table I, the traiig ad testig datasets are deoted as HM_TraiDS ad HM_TestDS, respectively. The umber of data istaces that are labeled as {Still}, {Up}, ad {Dow} are deoted as #Still, #Up, ad #Dow, respectively. To recogize head motios, the J48 decisio tree learig algorithm offered i Weka [27] was used. The classifier was traied usig the 48 features of the 3-axis accelerometer ad 3- axis gyroscope data, preseted i Sectio III.D. Fig. 12 shows the recogitio results of the costructed classifier for each head-motio type o the HM_TestDS dataset. As show i Fig. 12, the classifier achieves a high recogitio rate, with a average precisio of 89.1%, a average recall of 88%, ad a average F-measure of 87.8%. B. Performace of Pe-Motio Module The performace of the Pe-Motio module was assessed i the same classroom eviromet. Three pe motios were cosidered: {Still, Write, Other}. The participats were asked to use the desiged smart pe, equipped with a 3-axis accelerometer ad a 3-axis gyroscope. A video camera was used to record the sessio for groud truth labelig. Totally, data istaces were collected, where data istaces were used for the traiig phase ad the other for the testig phase. Characteristics of the collected traiig ad testig datasets for pe-motio recogitio are show i Table II. I Table II, the traiig ad testig datasets are deoted as PM_TraiDS ad PM_TestDS, respectively. The umber of data istaces that are labeled as {Still}, {Write}, ad {Other} are deoted as #Still, #Write, ad #Other, respectively. A J48 classifier was traied usig the 48 features preseted i sectio III.D. The precisio, recall ad F-Measure were used to assess the recogitio rate of the pe-motio module. Fig. 13 shows the recogitio results of the costructed pe-motio classifier for each pe-motio type o the PM_TestDS dataset. As show i Fig. 13, the costructed classifier achieves remarkable recogitio rate with a average precisio of 97.27%, a average recall of 97.6%, ad a average F- measure of 97.6%. TABLE I. TABLE II. CHARATERISTICS OF TRANING AND TESTING DATASETS FOR THE HEAD-MOTION MODULE. Dataset #Istace #Sill #Up #Dow HM_TraiDS HM_TestDS CHARATERISTICS OF TRANING AND TESTING DATASETS FOR THE PEN-MOTION MODULE. Dataset #Istace #Sill #Write #Other PM_TraiDS PM_TestDS Precisio Recall F-Measure Still Up Dow Fig. 12. Effectiveess of the head-motio module o the HM_TestDS dataset. Precisio Recall F-Measure Still Write Other Fig. 13. Effectiveess of the pe-motio module o the PM_TestDS dataset M 2M_Oblique 2.5M Miimum Matchig Number Threshold Fig. 14. Accuracy of the slide detectio module for room EC5B whe the matchig umber threshold is varied M 4M_Oblique 5.5M Miimum Matchig Number Threshold Fig. 15. Accuracy of the slide detectio module for room EC543 whe the matchig umber threshold is varied.

8 Accuracy (%) Accuracy (%) C. Performace of Slide Detectio Module The performace of the Slide Detectio module was assessed i a experimet coducted i two differet locatios: a 1-perso room (EC5B) ad a -perso room (EC543) at Natioal Chiao Tug Uiversity. I this experimet, five participats were asked to atted a lecture ad press a butto whe they were lookig at the slides, to collect groud truth labels for the recorded data. The wearable device cameras captured images with a resolutio. The miimum offceter threshold was set to 3. We evaluated the accuracy of the Slide Detectio module uder the followig three factors: (1) the miimum matchig umber threshold, (2) the distace betwee the user ad projectio scree, ad (3) the viewig agle (i.e., oblique or frot facig) of the user. Fig. 14 ad Fig. 15 show the results. Based o Fig. 14 ad 15, we make the followig observatios. First, we observe that if the miimum matchig umber threshold is icreased, the accuracy icreases util it reaches a peak, ad the the accuracy decreases afterwards. This is reasoable sice a higher threshold meas a stricter requiremet for image recogitio. Whe the threshold is set too low, oise images may be cosidered as matchig with the slide logo. Secod, the accuracy decreases as the distace betwee the user ad the projectio scree icreases. This is because as the distace icreases, the captured image of the logo become smaller ad less clear. As a result, the recogitio rate of the system decreases. Third, the system is more accurate for frot facig participats that for those viewig the projectio scree from a oblique perspective. D. Performace of Attetio Iferece Egie The performace of the proposed attetio iferece egie, usig the rule-based approach or the data-drive approach, was also evaluated. The experimet was coducted with te participats. Each participat atteded a distict 5- miute lecture i a classroom, seated frot facig, at a 2m distace from the projectio scree. A video camera was used for groud truth labelig. The head-motio ad pe-motio classifiers, itroduced i Sectio IV.A ad Sectio IV.B, were used for head-motio ad pe-motio recogitios, respectively. For the Slide Detectio module, the miimum matchig umber threshold ad the miimum off-ceter threshold θ were set to 18 ad 3, respectively. For the Had Detectio module, the miimum pixel threshold was set to,. Fig. 16 shows results obtaied by the rule-based approach. It was able to correctly idetify the attetio state of beig focused ad ufocused with a average precisio of 78.1%, a average recall of 61%, ad a average F-measure of 69.2%. Fig. 17 shows the accuracy of the datadrive approach usig various machie learig algorithms, icludig J48 decisio tree, Radom Forest [28], ad Support Vector Machie (SVM) [28]. As show i this figure, differet classifiers have similar accuracy, but the J48 decisio tree achieves the highest accuracy. Fig. 18 shows the performace of the data-drive approach usig J48 i terms of precisio, recall, ad F-measure. The proposed data-drive approach achieves aroud % F-measure values for both the Focused ad Ufocused attetio states, which demostrates that it is effective at idetifyig the attetio states of studets i class. Precisio Recall F-measure Focused Ufocused Fig. 16. The performace of rule-based approach i terms of precisio, recall ad F-measure. J.48 RadomForest SVM J.48 RadomForest SVM Fig. 17. The performace comparisio of the data-drive approach usig differet classifiers. Focused Ufocused Fig. 18. The performace of data-drive approach i terms of precisio, recall ad F-measure. Rule-based Approach Precisio Recall F-measure Data-drive Approach Fig. 19. The performace compariso of two attetio iferece approaches. Fig. 19 compares the accuracy of the rule-based approach ad the data-drive approach. It is foud that the data-drive approach is the most accurate. This is because the data-drive approach uses machie learig methods to fid hidde relatioships betwee extracted high-level features ad attetio states for attetio iferece, while the rule-based approach uses ituitive decisio rules for iferece. However, the rule-based approach ca output the recogitio result every three secods, while the data-drive approach eeds to collect thirty secods of data to produce a result.

9 V. CONCLUSIONS AND FUTURE WORKS Assessig attetio levels of studets is highly desirable. It ca let studets uderstad their ow learig behavior so that they ca lear more efficietly. For teachers, this iformatio is also very importat as it idicates how studets react to their teachig. This iformatio is thus crucial for the desig of strategies for capturig ad maitaiig studet s attetio. I this work, we have demostrated that it is possible to accurately ifer the attetio levels of studets i a classroom based o their exteral behaviors, usig various types of wearable sesors. A ew wearable system was desiged, cosistig of four modules, amed the Head-Motio, Pe- Motio, Visual-Focus, ad Active-App modules. These modules recogize ad track differet activities of users, to provide behavioral data to the desiged attetio iferece egie, which calculates users attetio levels. This egie is equipped with two ovel attetio iferece algorithms, amed the rule-based approach ad the data-drive approach. The former uses ituitive decisio rules to ifer the attetio state of a studet, while the latter relies o machie learig methods. The egie ca geerate visual reports to idicate each studet s attetio levels ad class-related activities. Extesive experimets were coducted to evaluate the proposed system. Results have show that it is highly accurate i various real-life settigs. The proposed methodology thus has the potetial of greatly icreasig learig ad teachig efficiecy i the classroom. For future work, other types of activities ad sesors will be cosidered to further refie the attetio iferece methods preseted i this paper. Moreover, a larger scale evaluatio of the system is also plaed, where full classrooms of studets will be equipped with the desiged wearable modules. ACKNOWLEDGEMENT This work was partially supported by Miistry of Sciece ad Techology, Taiwa, R.O.C. uder grat E MY3. REFERENCES [1] J. R. Aderso, Cogitive Psychology ad Its Implicatios,, 6th ed. Worth Publishers, 4. [2] A. Al-Fuqaha, M. Guizai, M. Mohammadi, M. Aledhari, ad M. Ayyash, Iteret of Thigs: A Survey o Eablig Techologies, Protocols, ad Applicatios, IEEE Commuicatios Surveys & Tutorials, Vol. 17,. 4, pp , 15. [3] S. Asteriadis, K. Karpouzis, ad S. Kollias, The Importace of Eye Gaze ad Head Pose to Estimatig Levels of Attetio, i Proc. of IEEE Iteratioal Coferece o Games ad Virtual Worlds for Serious Applicatios, pp , 11. [4] H. Bay, A. Ess, T. Tuytelaarsm ad L. V. Gool, Speeded-Up Robust Features (SURF), Computer Visio ad Image Uderstadig, Vol. 11, Issue 3, pp , 8. [5] L. Bao ad S. S. Itille, Activity Recogitio from User-aotated Acceleratio Data, i Proc. of Iteratioal Coferece o Pervasive, pp. 1-17, 4. [6] C. Che, R. Jafari, ad N. Kehtaravaz, Improvig Huma Actio Recogitio usig Fusio of Depth Camera ad Iertial Sesors, IEEE Tras. o Huma-Machie Systems, Vol. 45,. 1, pp , 15. [7] DOCUMENTATION OpeCV, [8] H. Hug, G. Eglebiee, ad J. Kools, Classifyig Social Actios with a Sigle Sccelerometer, i Proc. of ACM Iteratioal Joit Coferece o Pervasive ad Ubiquitous Computig, pp. 7-21, 13. [9] T. Huyh, M. Fritz, ad B. Schiele, Discovery of Activity Patters usig Topic Models, i Proc. of ACM Iteratioal Coferece o Ubiquitous Computig, pp. 1-19, 8. [1] Iteratioal Data Corporatio (IDC), [11] A. M. Kha, Y.-K. Lee, S. Y. Lee, ad T.-S. Kim, A Triaxial Accelerometer-based Physical-activity Recogitio via Augmetedsigal Features ad a Hierarchical Recogizer, IEEE Tras. o Iformatio Techology i Biomedicie, Vol. 14,.5, pp , 1. [12] P. Kakumau, S. Makrogiais, ad N. Bourbakis, A Survey of Skicolor Modelig ad Detectio Methods, Patter Recogitio, Vol., Issue 3, pp , 7. [13] D. M. Laders, S. H. Boutcher, ad M. Q. Wag, A Psychophysiological Study of Archery Performace, Research Quarterly for Exercise ad Sport, Vol. 57, pp , [14] G. W. Lauth, B. G. Heubeck, ad K. Mackowiak, Observatio of Childre with Attetio-Deficit Hyperactivity (ADHD) problems i Three Natural Classroom Cotexts, British Joural of Educatioal Psychology, Vol. 76, pp , 6. [15] Y. Li, X. Li, M. Ratcliffe, L. Liu, Y. Qi, ad Q. Liu, A Real-time EEG- Based BCI System for Attetio Recogitio i Ubiquitous Eviromet, i Proc. of ACM Iteratioal Workshop o Ubiquitous Affective Awareess ad Itelliget Iteractio, pp. 33-, 11. [16] Oscar D. Lara ad Miguel A. Labrador, A Survey o Huma Activity Recogitio Usig Wearable Sesors, IEEE Commuicatios Surveys & Tutorials, Vol. 15,. 3, 13. [17] D. Mig, Y. Xi, M. Zhag, H. Qi, L. Cheg, B. Wa, ad L. Li, Electroecephalograph (EEG) Sigal Processig Method of Motor Imagiary Potetial for Attetio Level Classificatio, i Proc. of IEEE Iteratioal Coferece o Egieerig i Medicie ad Biology Society, pp , 9. [18] V. D. Nguye, M. T. Le, A. D. Do, H. H. Duog, T. D. Thai, ad D. H. Tra, A Efficiet Camera-based Surveillace for Fall Detectio of Elderly People, i Proc. of IEEE Coferece o Idustrial Electroics ad Applicatios, pp , 14. [19] W. R. Pruehser ad J. D. Ederie, Ifrared Radiat Itesity Exposure Safety Study for the Eye Tracker, Biomed Sci Istrum, Vol. 41, pp , 5. [] Raspberry Pi Foudatio, [21] S. Rallapalli, A. Gaesa, K. K. Chitalapudi, V. N. Padmaabha, ad L. Qiu, Eablig Physical Aalytics i Retail Stores usig Smart Glasses, i Proc. of ACM Iteratioal Coferece o Mobile Computig ad Networkig, pp , 14. [22] J. M. Schepers, The Costructio ad Evaluatio of a Attetio Questioaire, SA Joural of Idustrial Psychology, Vol. 33, pp , 7. [23] M. M. Sohlberg ad C. A. Mateer, Effectiveess of a Attetio Traiig Program, Joural of Cliical ad Experimetal Neuropsychology, Vol. 9, Issue 2, pp , [24] T. Shay, S. J. Redmod, M. R. Narayaa, ad N. H. Lovell, Sesorsbased Wearable Systems for Moitorig of Huma Movemet ad Falls, IEEE Sesors Joural, Vol. 12, pp , 12. [25] The Apache Software Foudatio, [26] The SesorTag Story-Texas Istrumets, SesorTag&HQS=sesortag. [27] Weka 3: Data Miig Software i Java, [28] I. Witte, E. Frak, ad M. A. Hall, Data Miig: Practical Machie Learig Tools ad Techiques, 3rd ed. Morga Kaufma, 5. [29] Y. Xu, N. Stojaovic, L. Stojaovic, ad T. Schuchert, Efficiet Huma Attetio Detectio based o Itelliget Complex Evet Processig, i Proc. of ACM Iteratioal Coferece o Distributed Evet-Based Systems, pp , 12. [3] Z. Ye, Y. Li, A. Fathi, Y. Ha, A. Rozga, G. D. Abowd ad, J. M. Rehg, Detectig Eye Cotact usig Wearable Eye-trackig Glasses, i Proc. of ACM Iteratioal Coferece o Ubiquitous Computig, pp , 12.

Improving Template Based Spike Detection

Improving Template Based Spike Detection Improvig Template Based Spike Detectio Kirk Smith, Member - IEEE Portlad State Uiversity petra@ee.pdx.edu Abstract Template matchig algorithms like SSE, Covolutio ad Maximum Likelihood are well kow for

More information

Neuro Fuzzy Model for Human Face Expression Recognition

Neuro Fuzzy Model for Human Face Expression Recognition IOSR Joural of Computer Egieerig (IOSRJCE) ISSN : 2278-0661 Volume 1, Issue 2 (May-Jue 2012), PP 01-06 Neuro Fuzzy Model for Huma Face Expressio Recogitio Mr. Mayur S. Burage 1, Prof. S. V. Dhopte 2 1

More information

Euclidean Distance Based Feature Selection for Fault Detection Prediction Model in Semiconductor Manufacturing Process

Euclidean Distance Based Feature Selection for Fault Detection Prediction Model in Semiconductor Manufacturing Process Vol.133 (Iformatio Techology ad Computer Sciece 016), pp.85-89 http://dx.doi.org/10.1457/astl.016. Euclidea Distace Based Feature Selectio for Fault Detectio Predictio Model i Semicoductor Maufacturig

More information

Sectio 4, a prototype project of settig field weight with AHP method is developed ad the experimetal results are aalyzed. Fially, we coclude our work

Sectio 4, a prototype project of settig field weight with AHP method is developed ad the experimetal results are aalyzed. Fially, we coclude our work 200 2d Iteratioal Coferece o Iformatio ad Multimedia Techology (ICIMT 200) IPCSIT vol. 42 (202) (202) IACSIT Press, Sigapore DOI: 0.7763/IPCSIT.202.V42.0 Idex Weight Decisio Based o AHP for Iformatio Retrieval

More information

3D Model Retrieval Method Based on Sample Prediction

3D Model Retrieval Method Based on Sample Prediction 20 Iteratioal Coferece o Computer Commuicatio ad Maagemet Proc.of CSIT vol.5 (20) (20) IACSIT Press, Sigapore 3D Model Retrieval Method Based o Sample Predictio Qigche Zhag, Ya Tag* School of Computer

More information

Mobile terminal 3D image reconstruction program development based on Android Lin Qinhua

Mobile terminal 3D image reconstruction program development based on Android Lin Qinhua Iteratioal Coferece o Automatio, Mechaical Cotrol ad Computatioal Egieerig (AMCCE 05) Mobile termial 3D image recostructio program developmet based o Adroid Li Qihua Sichua Iformatio Techology College

More information

A Novel Feature Extraction Algorithm for Haar Local Binary Pattern Texture Based on Human Vision System

A Novel Feature Extraction Algorithm for Haar Local Binary Pattern Texture Based on Human Vision System A Novel Feature Extractio Algorithm for Haar Local Biary Patter Texture Based o Huma Visio System Liu Tao 1,* 1 Departmet of Electroic Egieerig Shaaxi Eergy Istitute Xiayag, Shaaxi, Chia Abstract The locality

More information

Image Segmentation EEE 508

Image Segmentation EEE 508 Image Segmetatio Objective: to determie (etract) object boudaries. It is a process of partitioig a image ito distict regios by groupig together eighborig piels based o some predefied similarity criterio.

More information

( n+1 2 ) , position=(7+1)/2 =4,(median is observation #4) Median=10lb

( n+1 2 ) , position=(7+1)/2 =4,(median is observation #4) Median=10lb Chapter 3 Descriptive Measures Measures of Ceter (Cetral Tedecy) These measures will tell us where is the ceter of our data or where most typical value of a data set lies Mode the value that occurs most

More information

Optimization for framework design of new product introduction management system Ma Ying, Wu Hongcui

Optimization for framework design of new product introduction management system Ma Ying, Wu Hongcui 2d Iteratioal Coferece o Electrical, Computer Egieerig ad Electroics (ICECEE 2015) Optimizatio for framework desig of ew product itroductio maagemet system Ma Yig, Wu Hogcui Tiaji Electroic Iformatio Vocatioal

More information

Chapter 1. Introduction to Computers and C++ Programming. Copyright 2015 Pearson Education, Ltd.. All rights reserved.

Chapter 1. Introduction to Computers and C++ Programming. Copyright 2015 Pearson Education, Ltd.. All rights reserved. Chapter 1 Itroductio to Computers ad C++ Programmig Copyright 2015 Pearso Educatio, Ltd.. All rights reserved. Overview 1.1 Computer Systems 1.2 Programmig ad Problem Solvig 1.3 Itroductio to C++ 1.4 Testig

More information

Evaluation scheme for Tracking in AMI

Evaluation scheme for Tracking in AMI A M I C o m m u i c a t i o A U G M E N T E D M U L T I - P A R T Y I N T E R A C T I O N http://www.amiproject.org/ Evaluatio scheme for Trackig i AMI S. Schreiber a D. Gatica-Perez b AMI WP4 Trackig:

More information

IMP: Superposer Integrated Morphometrics Package Superposition Tool

IMP: Superposer Integrated Morphometrics Package Superposition Tool IMP: Superposer Itegrated Morphometrics Package Superpositio Tool Programmig by: David Lieber ( 03) Caisius College 200 Mai St. Buffalo, NY 4208 Cocept by: H. David Sheets, Dept. of Physics, Caisius College

More information

Harris Corner Detection Algorithm at Sub-pixel Level and Its Application Yuanfeng Han a, Peijiang Chen b * and Tian Meng c

Harris Corner Detection Algorithm at Sub-pixel Level and Its Application Yuanfeng Han a, Peijiang Chen b * and Tian Meng c Iteratioal Coferece o Computatioal Sciece ad Egieerig (ICCSE 015) Harris Corer Detectio Algorithm at Sub-pixel Level ad Its Applicatio Yuafeg Ha a, Peijiag Che b * ad Tia Meg c School of Automobile, Liyi

More information

Ontology-based Decision Support System with Analytic Hierarchy Process for Tour Package Selection

Ontology-based Decision Support System with Analytic Hierarchy Process for Tour Package Selection 2017 Asia-Pacific Egieerig ad Techology Coferece (APETC 2017) ISBN: 978-1-60595-443-1 Otology-based Decisio Support System with Aalytic Hierarchy Process for Tour Pacage Selectio Tie-We Sug, Chia-Jug Lee,

More information

Task scenarios Outline. Scenarios in Knowledge Extraction. Proposed Framework for Scenario to Design Diagram Transformation

Task scenarios Outline. Scenarios in Knowledge Extraction. Proposed Framework for Scenario to Design Diagram Transformation 6-0-0 Kowledge Trasformatio from Task Scearios to View-based Desig Diagrams Nima Dezhkam Kamra Sartipi {dezhka, sartipi}@mcmaster.ca Departmet of Computig ad Software McMaster Uiversity CANADA SEKE 08

More information

Application of Decision Tree and Support Vector Machine for Inspecting Bubble Defects on LED Sealing Glue Images

Application of Decision Tree and Support Vector Machine for Inspecting Bubble Defects on LED Sealing Glue Images 66 Applicatio of Decisio Tree ad Support Vector Machie for Ispectig Bubble Defects o LED Sealig Glue Images * Chua-Yu Chag ad Yi-Feg Li Abstract Bubble defect ispectio is a importat step i light-emittig

More information

Stone Images Retrieval Based on Color Histogram

Stone Images Retrieval Based on Color Histogram Stoe Images Retrieval Based o Color Histogram Qiag Zhao, Jie Yag, Jigyi Yag, Hogxig Liu School of Iformatio Egieerig, Wuha Uiversity of Techology Wuha, Chia Abstract Stoe images color features are chose

More information

Dynamic Programming and Curve Fitting Based Road Boundary Detection

Dynamic Programming and Curve Fitting Based Road Boundary Detection Dyamic Programmig ad Curve Fittig Based Road Boudary Detectio SHYAM PRASAD ADHIKARI, HYONGSUK KIM, Divisio of Electroics ad Iformatio Egieerig Chobuk Natioal Uiversity 664-4 Ga Deokji-Dog Jeoju-City Jeobuk

More information

Pattern Recognition Systems Lab 1 Least Mean Squares

Pattern Recognition Systems Lab 1 Least Mean Squares Patter Recogitio Systems Lab 1 Least Mea Squares 1. Objectives This laboratory work itroduces the OpeCV-based framework used throughout the course. I this assigmet a lie is fitted to a set of poits usig

More information

. Written in factored form it is easy to see that the roots are 2, 2, i,

. Written in factored form it is easy to see that the roots are 2, 2, i, CMPS A Itroductio to Programmig Programmig Assigmet 4 I this assigmet you will write a java program that determies the real roots of a polyomial that lie withi a specified rage. Recall that the roots (or

More information

Low Complexity H.265/HEVC Coding Unit Size Decision for a Videoconferencing System

Low Complexity H.265/HEVC Coding Unit Size Decision for a Videoconferencing System BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 15, No 6 Special Issue o Logistics, Iformatics ad Service Sciece Sofia 2015 Prit ISSN: 1311-9702; Olie ISSN: 1314-4081 DOI:

More information

A New Network-based Algorithm for Human Activity Recognition in Videos

A New Network-based Algorithm for Human Activity Recognition in Videos IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, A New Network-based Algorithm for Huma Activity Recogitio i Videos Weiyao Li, Yuazhe Che, Jiaxi Wu, Hali Wag, Bi Sheg, ad Hogxiag Li Abstract

More information

Analysis of Server Resource Consumption of Meteorological Satellite Application System Based on Contour Curve

Analysis of Server Resource Consumption of Meteorological Satellite Application System Based on Contour Curve Advaces i Computer, Sigals ad Systems (2018) 2: 19-25 Clausius Scietific Press, Caada Aalysis of Server Resource Cosumptio of Meteorological Satellite Applicatio System Based o Cotour Curve Xiagag Zhao

More information

Practical Experience Recording and Indexing of Life Log Video

Practical Experience Recording and Indexing of Life Log Video Practical Experiece Recordig ad Idexig of Life Log Video Datchakor Tacharoe Dept. of Electroic Egieerig The Uiversity of Tokyo 5--5 Kashiwaoha, Kashiwa, Chiba, 277-856 JAPAN Phoe: (+8)-4-736-3888 dta@hal.k.u-tokyo.ac.jp

More information

DATA MINING II - 1DL460

DATA MINING II - 1DL460 DATA MINING II - 1DL460 Sprig 2017 A secod course i data miig http://www.it.uu.se/edu/course/homepage/ifoutv2/vt17/ Kjell Orsbor Uppsala Database Laboratory Departmet of Iformatio Techology, Uppsala Uiversity,

More information

Accuracy Improvement in Camera Calibration

Accuracy Improvement in Camera Calibration Accuracy Improvemet i Camera Calibratio FaJie L Qi Zag ad Reihard Klette CITR, Computer Sciece Departmet The Uiversity of Aucklad Tamaki Campus, Aucklad, New Zealad fli006, qza001@ec.aucklad.ac.z r.klette@aucklad.ac.z

More information

Python Programming: An Introduction to Computer Science

Python Programming: An Introduction to Computer Science Pytho Programmig: A Itroductio to Computer Sciece Chapter 1 Computers ad Programs 1 Objectives To uderstad the respective roles of hardware ad software i a computig system. To lear what computer scietists

More information

Pruning and Summarizing the Discovered Time Series Association Rules from Mechanical Sensor Data Qing YANG1,a,*, Shao-Yu WANG1,b, Ting-Ting ZHANG2,c

Pruning and Summarizing the Discovered Time Series Association Rules from Mechanical Sensor Data Qing YANG1,a,*, Shao-Yu WANG1,b, Ting-Ting ZHANG2,c Advaces i Egieerig Research (AER), volume 131 3rd Aual Iteratioal Coferece o Electroics, Electrical Egieerig ad Iformatio Sciece (EEEIS 2017) Pruig ad Summarizig the Discovered Time Series Associatio Rules

More information

Empirical Validate C&K Suite for Predict Fault-Proneness of Object-Oriented Classes Developed Using Fuzzy Logic.

Empirical Validate C&K Suite for Predict Fault-Proneness of Object-Oriented Classes Developed Using Fuzzy Logic. Empirical Validate C&K Suite for Predict Fault-Proeess of Object-Orieted Classes Developed Usig Fuzzy Logic. Mohammad Amro 1, Moataz Ahmed 1, Kaaa Faisal 2 1 Iformatio ad Computer Sciece Departmet, Kig

More information

Title: Robust Registration of Multimodal Remote Sensing Images Based on Structural Similarity.

Title: Robust Registration of Multimodal Remote Sensing Images Based on Structural Similarity. 7 IEEE. Persoal use of this material is permitted. Permissio from IEEE must be obtaied for all other uses, i ay curret or future media, icludig repritig/republishig this material for advertisig or promotioal

More information

Designing a learning system

Designing a learning system CS 75 Machie Learig Lecture Desigig a learig system Milos Hauskrecht milos@cs.pitt.edu 539 Seott Square, x-5 people.cs.pitt.edu/~milos/courses/cs75/ Admiistrivia No homework assigmet this week Please try

More information

A SOFTWARE MODEL FOR THE MULTILAYER PERCEPTRON

A SOFTWARE MODEL FOR THE MULTILAYER PERCEPTRON A SOFTWARE MODEL FOR THE MULTILAYER PERCEPTRON Roberto Lopez ad Eugeio Oñate Iteratioal Ceter for Numerical Methods i Egieerig (CIMNE) Edificio C1, Gra Capitá s/, 08034 Barceloa, Spai ABSTRACT I this work

More information

Term Project Report. This component works to detect gesture from the patient as a sign of emergency message and send it to the emergency manager.

Term Project Report. This component works to detect gesture from the patient as a sign of emergency message and send it to the emergency manager. CS2310 Fial Project Loghao Li Term Project Report Itroductio I this project, I worked o expadig exercise 4. What I focused o is makig the real gesture recogizig sesor ad desig proper gestures ad recogizig

More information

New HSL Distance Based Colour Clustering Algorithm

New HSL Distance Based Colour Clustering Algorithm The 4th Midwest Artificial Itelligece ad Cogitive Scieces Coferece (MAICS 03 pp 85-9 New Albay Idiaa USA April 3-4 03 New HSL Distace Based Colour Clusterig Algorithm Vasile Patrascu Departemet of Iformatics

More information

Intermediate Statistics

Intermediate Statistics Gait Learig Guides Itermediate Statistics Data processig & display, Cetral tedecy Author: Raghu M.D. STATISTICS DATA PROCESSING AND DISPLAY Statistics is the study of data or umerical facts of differet

More information

IMAGE-BASED MODELING AND RENDERING 1. HISTOGRAM AND GMM. I-Chen Lin, Dept. of CS, National Chiao Tung University

IMAGE-BASED MODELING AND RENDERING 1. HISTOGRAM AND GMM. I-Chen Lin, Dept. of CS, National Chiao Tung University IMAGE-BASED MODELING AND RENDERING. HISTOGRAM AND GMM I-Che Li, Dept. of CS, Natioal Chiao Tug Uiversity Outlie What s the itesity/color histogram? What s the Gaussia Mixture Model (GMM? Their applicatios

More information

Performance Plus Software Parameter Definitions

Performance Plus Software Parameter Definitions Performace Plus+ Software Parameter Defiitios/ Performace Plus Software Parameter Defiitios Chapma Techical Note-TG-5 paramete.doc ev-0-03 Performace Plus+ Software Parameter Defiitios/2 Backgroud ad Defiitios

More information

Outline. Research Definition. Motivation. Foundation of Reverse Engineering. Dynamic Analysis and Design Pattern Detection in Java Programs

Outline. Research Definition. Motivation. Foundation of Reverse Engineering. Dynamic Analysis and Design Pattern Detection in Java Programs Dyamic Aalysis ad Desig Patter Detectio i Java Programs Outlie Lei Hu Kamra Sartipi {hul4, sartipi}@mcmasterca Departmet of Computig ad Software McMaster Uiversity Caada Motivatio Research Problem Defiitio

More information

Adaptive Resource Allocation for Electric Environmental Pollution through the Control Network

Adaptive Resource Allocation for Electric Environmental Pollution through the Control Network Available olie at www.sciecedirect.com Eergy Procedia 6 (202) 60 64 202 Iteratioal Coferece o Future Eergy, Eviromet, ad Materials Adaptive Resource Allocatio for Electric Evirometal Pollutio through the

More information

Data diverse software fault tolerance techniques

Data diverse software fault tolerance techniques Data diverse software fault tolerace techiques Complemets desig diversity by compesatig for desig diversity s s limitatios Ivolves obtaiig a related set of poits i the program data space, executig the

More information

arxiv: v2 [cs.ds] 24 Mar 2018

arxiv: v2 [cs.ds] 24 Mar 2018 Similar Elemets ad Metric Labelig o Complete Graphs arxiv:1803.08037v [cs.ds] 4 Mar 018 Pedro F. Felzeszwalb Brow Uiversity Providece, RI, USA pff@brow.edu March 8, 018 We cosider a problem that ivolves

More information

APPLICATION NOTE PACE1750AE BUILT-IN FUNCTIONS

APPLICATION NOTE PACE1750AE BUILT-IN FUNCTIONS APPLICATION NOTE PACE175AE BUILT-IN UNCTIONS About This Note This applicatio brief is iteded to explai ad demostrate the use of the special fuctios that are built ito the PACE175AE processor. These powerful

More information

Mining from Quantitative Data with Linguistic Minimum Supports and Confidences

Mining from Quantitative Data with Linguistic Minimum Supports and Confidences Miig from Quatitative Data with Liguistic Miimum Supports ad Cofideces Tzug-Pei Hog, Mig-Jer Chiag ad Shyue-Liag Wag Departmet of Electrical Egieerig Natioal Uiversity of Kaohsiug Kaohsiug, 8, Taiwa, R.O.C.

More information

A New Morphological 3D Shape Decomposition: Grayscale Interframe Interpolation Method

A New Morphological 3D Shape Decomposition: Grayscale Interframe Interpolation Method A ew Morphological 3D Shape Decompositio: Grayscale Iterframe Iterpolatio Method D.. Vizireau Politehica Uiversity Bucharest, Romaia ae@comm.pub.ro R. M. Udrea Politehica Uiversity Bucharest, Romaia mihea@comm.pub.ro

More information

Global Support Guide. Verizon WIreless. For the BlackBerry 8830 World Edition Smartphone and the Motorola Z6c

Global Support Guide. Verizon WIreless. For the BlackBerry 8830 World Edition Smartphone and the Motorola Z6c Verizo WIreless Global Support Guide For the BlackBerry 8830 World Editio Smartphoe ad the Motorola Z6c For complete iformatio o global services, please refer to verizowireless.com/vzglobal. Whether i

More information

Improvement of the Orthogonal Code Convolution Capabilities Using FPGA Implementation

Improvement of the Orthogonal Code Convolution Capabilities Using FPGA Implementation Improvemet of the Orthogoal Code Covolutio Capabilities Usig FPGA Implemetatio Naima Kaabouch, Member, IEEE, Apara Dhirde, Member, IEEE, Saleh Faruque, Member, IEEE Departmet of Electrical Egieerig, Uiversity

More information

Elementary Educational Computer

Elementary Educational Computer Chapter 5 Elemetary Educatioal Computer. Geeral structure of the Elemetary Educatioal Computer (EEC) The EEC coforms to the 5 uits structure defied by vo Neuma's model (.) All uits are preseted i a simplified

More information

Improving Face Recognition Rate by Combining Eigenface Approach and Case-based Reasoning

Improving Face Recognition Rate by Combining Eigenface Approach and Case-based Reasoning Improvig Face Recogitio Rate by Combiig Eigeface Approach ad Case-based Reasoig Haris Supic, ember, IAENG Abstract There are may approaches to the face recogitio. This paper presets a approach that combies

More information

The Closest Line to a Data Set in the Plane. David Gurney Southeastern Louisiana University Hammond, Louisiana

The Closest Line to a Data Set in the Plane. David Gurney Southeastern Louisiana University Hammond, Louisiana The Closest Lie to a Data Set i the Plae David Gurey Southeaster Louisiaa Uiversity Hammod, Louisiaa ABSTRACT This paper looks at three differet measures of distace betwee a lie ad a data set i the plae:

More information

Fire Recognition in Video. Walter Phillips III Mubarak Shah Niels da Vitoria Lobo.

Fire Recognition in Video. Walter Phillips III Mubarak Shah Niels da Vitoria Lobo. Fire Recogitio i Video Walter Phillips III Mubarak Shah Niels da Vitoria Lobo {wrp65547,shah,iels}@cs.ucf.edu Computer Visio Laboratory Departmet of Computer Sciece Uiversity of Cetral Florida Orlado,

More information

Appendix D. Controller Implementation

Appendix D. Controller Implementation COMPUTER ORGANIZATION AND DESIGN The Hardware/Software Iterface 5 th Editio Appedix D Cotroller Implemetatio Cotroller Implemetatios Combiatioal logic (sigle-cycle); Fiite state machie (multi-cycle, pipelied);

More information

Auto-recognition Method for Pointer-type Meter Based on Binocular Vision

Auto-recognition Method for Pointer-type Meter Based on Binocular Vision JOURNAL OF COMPUTERS, VOL. 9, NO. 4, APRIL 204 787 Auto-recogitio Method for Poiter-type Meter Based o Biocular Visio Biao Yag School of Istrumet Sciece ad Egieerig, Southeast Uiversity, Najig 20096, Chia

More information

Fundamentals of Media Processing. Shin'ichi Satoh Kazuya Kodama Hiroshi Mo Duy-Dinh Le

Fundamentals of Media Processing. Shin'ichi Satoh Kazuya Kodama Hiroshi Mo Duy-Dinh Le Fudametals of Media Processig Shi'ichi Satoh Kazuya Kodama Hiroshi Mo Duy-Dih Le Today's topics Noparametric Methods Parze Widow k-nearest Neighbor Estimatio Clusterig Techiques k-meas Agglomerative Hierarchical

More information

Lecture 28: Data Link Layer

Lecture 28: Data Link Layer Automatic Repeat Request (ARQ) 2. Go ack N ARQ Although the Stop ad Wait ARQ is very simple, you ca easily show that it has very the low efficiecy. The low efficiecy comes from the fact that the trasmittig

More information

WEBSITE STRUCTURE IMPROVEMENT USING ANT COLONY TECHNIQUE

WEBSITE STRUCTURE IMPROVEMENT USING ANT COLONY TECHNIQUE WEBSITE STRUCTURE IMPROVEMENT USING ANT COLONY TECHNIQUE Wiwik Aggraei 1, Agyl Ardi Rahmadi 1, Radityo Prasetyo Wibowo 1 1 Iformatio System Departmet, Faculty of Iformatio Techology, Istitut Tekologi Sepuluh

More information

Text Feature Selection based on Feature Dispersion Degree and Feature Concentration Degree

Text Feature Selection based on Feature Dispersion Degree and Feature Concentration Degree Available olie at www.ijpe-olie.com vol. 13, o. 7, November 017, pp. 1159-1164 DOI: 10.3940/ijpe.17.07.p19.11591164 Text Feature Selectio based o Feature Dispersio Degree ad Feature Cocetratio Degree Zhifeg

More information

Analysis Metrics. Intro to Algorithm Analysis. Slides. 12. Alg Analysis. 12. Alg Analysis

Analysis Metrics. Intro to Algorithm Analysis. Slides. 12. Alg Analysis. 12. Alg Analysis Itro to Algorithm Aalysis Aalysis Metrics Slides. Table of Cotets. Aalysis Metrics 3. Exact Aalysis Rules 4. Simple Summatio 5. Summatio Formulas 6. Order of Magitude 7. Big-O otatio 8. Big-O Theorems

More information

CSC 220: Computer Organization Unit 11 Basic Computer Organization and Design

CSC 220: Computer Organization Unit 11 Basic Computer Organization and Design College of Computer ad Iformatio Scieces Departmet of Computer Sciece CSC 220: Computer Orgaizatio Uit 11 Basic Computer Orgaizatio ad Desig 1 For the rest of the semester, we ll focus o computer architecture:

More information

Study on effective detection method for specific data of large database LI Jin-feng

Study on effective detection method for specific data of large database LI Jin-feng Iteratioal Coferece o Automatio, Mechaical Cotrol ad Computatioal Egieerig (AMCCE 205) Study o effective detectio method for specific data of large database LI Ji-feg (Vocatioal College of DogYig, Shadog

More information

Lecture Notes 6 Introduction to algorithm analysis CSS 501 Data Structures and Object-Oriented Programming

Lecture Notes 6 Introduction to algorithm analysis CSS 501 Data Structures and Object-Oriented Programming Lecture Notes 6 Itroductio to algorithm aalysis CSS 501 Data Structures ad Object-Orieted Programmig Readig for this lecture: Carrao, Chapter 10 To be covered i this lecture: Itroductio to algorithm aalysis

More information

Efficient Eye Location for Biomedical Imaging using Two-level Classifier Scheme

Efficient Eye Location for Biomedical Imaging using Two-level Classifier Scheme 828 Iteratioal Joural of Cotrol, Automatio, ad Systems, vol. 6, o. 6, pp. 828-835, December 2008 Efficiet Eye Locatio for Biomedical Imagig usig Two-level Classifier Scheme Mi Youg Nam, Xi Wag, ad Phill

More information

GE FUNDAMENTALS OF COMPUTING AND PROGRAMMING UNIT III

GE FUNDAMENTALS OF COMPUTING AND PROGRAMMING UNIT III GE2112 - FUNDAMENTALS OF COMPUTING AND PROGRAMMING UNIT III PROBLEM SOLVING AND OFFICE APPLICATION SOFTWARE Plaig the Computer Program Purpose Algorithm Flow Charts Pseudocode -Applicatio Software Packages-

More information

Lecture 5. Counting Sort / Radix Sort

Lecture 5. Counting Sort / Radix Sort Lecture 5. Coutig Sort / Radix Sort T. H. Corme, C. E. Leiserso ad R. L. Rivest Itroductio to Algorithms, 3rd Editio, MIT Press, 2009 Sugkyukwa Uiversity Hyuseug Choo choo@skku.edu Copyright 2000-2018

More information

Python Programming: An Introduction to Computer Science

Python Programming: An Introduction to Computer Science Pytho Programmig: A Itroductio to Computer Sciece Chapter 6 Defiig Fuctios Pytho Programmig, 2/e 1 Objectives To uderstad why programmers divide programs up ito sets of cooperatig fuctios. To be able to

More information

Fast Fourier Transform (FFT) Algorithms

Fast Fourier Transform (FFT) Algorithms Fast Fourier Trasform FFT Algorithms Relatio to the z-trasform elsewhere, ozero, z x z X x [ ] 2 ~ elsewhere,, ~ e j x X x x π j e z z X X π 2 ~ The DFS X represets evely spaced samples of the z- trasform

More information

Real-time Path Prediction and Grid-based Path Modeling Method Using GPS

Real-time Path Prediction and Grid-based Path Modeling Method Using GPS Iteratioal Joural of Applied Egieerig Research ISSN 0973-4562 Volume 12, Number 20 (2017) pp. 9997-10001 Research Idia Publicatios. http://www.ripublicatio.com Real-time Path Predictio ad Grid-based Path

More information

EFFECT OF QUERY FORMATION ON WEB SEARCH ENGINE RESULTS

EFFECT OF QUERY FORMATION ON WEB SEARCH ENGINE RESULTS Iteratioal Joural o Natural Laguage Computig (IJNLC) Vol. 2, No., February 203 EFFECT OF QUERY FORMATION ON WEB SEARCH ENGINE RESULTS Raj Kishor Bisht ad Ila Pat Bisht 2 Departmet of Computer Sciece &

More information

Evaluation of Support Vector Machine Kernels for Detecting Network Anomalies

Evaluation of Support Vector Machine Kernels for Detecting Network Anomalies Evaluatio of Support Vector Machie Kerels for Detectig Network Aomalies Prera Batta, Maider Sigh, Zhida Li, Qigye Dig, ad Ljiljaa Trajković Commuicatio Networks Laboratory http://www.esc.sfu.ca/~ljilja/cl/

More information

Identification of the Swiss Z24 Highway Bridge by Frequency Domain Decomposition Brincker, Rune; Andersen, P.

Identification of the Swiss Z24 Highway Bridge by Frequency Domain Decomposition Brincker, Rune; Andersen, P. Aalborg Uiversitet Idetificatio of the Swiss Z24 Highway Bridge by Frequecy Domai Decompositio Bricker, Rue; Aderse, P. Published i: Proceedigs of IMAC 2 Publicatio date: 22 Documet Versio Publisher's

More information

9.1. Sequences and Series. Sequences. What you should learn. Why you should learn it. Definition of Sequence

9.1. Sequences and Series. Sequences. What you should learn. Why you should learn it. Definition of Sequence _9.qxd // : AM Page Chapter 9 Sequeces, Series, ad Probability 9. Sequeces ad Series What you should lear Use sequece otatio to write the terms of sequeces. Use factorial otatio. Use summatio otatio to

More information

Ones Assignment Method for Solving Traveling Salesman Problem

Ones Assignment Method for Solving Traveling Salesman Problem Joural of mathematics ad computer sciece 0 (0), 58-65 Oes Assigmet Method for Solvig Travelig Salesma Problem Hadi Basirzadeh Departmet of Mathematics, Shahid Chamra Uiversity, Ahvaz, Ira Article history:

More information

Design and Implementation of Web Usage Mining Intelligent System in the Field of e-commerce

Design and Implementation of Web Usage Mining Intelligent System in the Field of e-commerce Available olie at www.sciecedirect.com Procedia Egieerig 30 (2012) 20 27 Iteratioal Coferece o Commuicatio Techology ad System Desig Desig ad Implemetatio of Web Usage Miig Itelliget System i the Field

More information

Computer Systems - HS

Computer Systems - HS What have we leared so far? Computer Systems High Level ENGG1203 2d Semester, 2017-18 Applicatios Sigals Systems & Cotrol Systems Computer & Embedded Systems Digital Logic Combiatioal Logic Sequetial Logic

More information

UNIT 4 Section 8 Estimating Population Parameters using Confidence Intervals

UNIT 4 Section 8 Estimating Population Parameters using Confidence Intervals UNIT 4 Sectio 8 Estimatig Populatio Parameters usig Cofidece Itervals To make ifereces about a populatio that caot be surveyed etirely, sample statistics ca be take from a SRS of the populatio ad used

More information

On-line cursive letter recognition using sequences of local minima/maxima. Robert Powalka

On-line cursive letter recognition using sequences of local minima/maxima. Robert Powalka O-lie cursive letter recogitio usig sequeces of local miima/maxima Summary Robert Powalka 19 th August 1993 This report presets the desig ad implemetatio of a o-lie cursive letter recogizer usig sequeces

More information

Bayesian Network Structure Learning from Attribute Uncertain Data

Bayesian Network Structure Learning from Attribute Uncertain Data Bayesia Network Structure Learig from Attribute Ucertai Data Wetig Sog 1,2, Jeffrey Xu Yu 3, Hog Cheg 3, Hogya Liu 4, Ju He 1,2,*, ad Xiaoyog Du 1,2 1 Key Labs of Data Egieerig ad Kowledge Egieerig, Miistry

More information

Lip Contour Extraction Based on Support Vector Machine

Lip Contour Extraction Based on Support Vector Machine Lip Cotour Extractio Based o Support Vector Machie Author Pa, Xiaosheg, Kog, Jiagpig, Liew, Ala Wee-Chug Published 008 Coferece Title CISP 008 : Proceedigs, First Iteratioal Cogress o Image ad Sigal Processig

More information

Which movie we can suggest to Anne?

Which movie we can suggest to Anne? ECOLE CENTRALE SUPELEC MASTER DSBI DECISION MODELING TUTORIAL COLLABORATIVE FILTERING AS A MODEL OF GROUP DECISION-MAKING You kow that the low-tech way to get recommedatios for products, movies, or etertaiig

More information

Using the Keyboard. Using the Wireless Keyboard. > Using the Keyboard

Using the Keyboard. Using the Wireless Keyboard. > Using the Keyboard 1 A wireless keyboard is supplied with your computer. The wireless keyboard uses a stadard key arragemet with additioal keys that perform specific fuctios. Usig the Wireless Keyboard Two AA alkalie batteries

More information

An Image Retrieval Method Based on Hu Invariant Moment and Improved Annular Histogram

An Image Retrieval Method Based on Hu Invariant Moment and Improved Annular Histogram http://dx.doi.org/10.5755/j01.eee.0.4.6888 ELEKTROIKA IR ELEKTROTECHIKA ISS 139 115 VOL. 0 O. 4 014 A Image Retrieval Method Based o Hu Ivariat Momet ad Improved Aular Histogram F. Xiag 1 H. Yog 1 S. Dada

More information

Algorithms for Disk Covering Problems with the Most Points

Algorithms for Disk Covering Problems with the Most Points Algorithms for Disk Coverig Problems with the Most Poits Bi Xiao Departmet of Computig Hog Kog Polytechic Uiversity Hug Hom, Kowloo, Hog Kog csbxiao@comp.polyu.edu.hk Qigfeg Zhuge, Yi He, Zili Shao, Edwi

More information

are two specific neighboring points, F( x, y)

are two specific neighboring points, F( x, y) $33/,&$7,212)7+(6(/)$92,',1* 5$1'20:$/.12,6(5('8&7,21$/*25,7+0,17+(&2/285,0$*(6(*0(17$7,21 %RJGDQ602/.$+HQU\N3$/86'DPLDQ%(5(6.$ 6LOHVLDQ7HFKQLFDO8QLYHUVLW\'HSDUWPHQWRI&RPSXWHU6FLHQFH $NDGHPLFND*OLZLFH32/$1'

More information

ANN WHICH COVERS MLP AND RBF

ANN WHICH COVERS MLP AND RBF ANN WHICH COVERS MLP AND RBF Josef Boští, Jaromír Kual Faculty of Nuclear Scieces ad Physical Egieerig, CTU i Prague Departmet of Software Egieerig Abstract Two basic types of artificial eural etwors Multi

More information

FEATURE BASED RECOGNITION OF TRAFFIC VIDEO STREAMS FOR ONLINE ROUTE TRACING

FEATURE BASED RECOGNITION OF TRAFFIC VIDEO STREAMS FOR ONLINE ROUTE TRACING FEATURE BASED RECOGNITION OF TRAFFIC VIDEO STREAMS FOR ONLINE ROUTE TRACING Christoph Busch, Ralf Dörer, Christia Freytag, Heike Ziegler Frauhofer Istitute for Computer Graphics, Computer Graphics Ceter

More information

Arithmetic Sequences

Arithmetic Sequences . Arithmetic Sequeces COMMON CORE Learig Stadards HSF-IF.A. HSF-BF.A.1a HSF-BF.A. HSF-LE.A. Essetial Questio How ca you use a arithmetic sequece to describe a patter? A arithmetic sequece is a ordered

More information

Panel for Adobe Premiere Pro CC Partner Solution

Panel for Adobe Premiere Pro CC Partner Solution Pael for Adobe Premiere Pro CC Itegratio for more efficiecy The makes video editig simple, fast ad coveiet. The itegrated pael gives users immediate access to all medialoopster features iside Adobe Premiere

More information

Bayesian approach to reliability modelling for a probability of failure on demand parameter

Bayesian approach to reliability modelling for a probability of failure on demand parameter Bayesia approach to reliability modellig for a probability of failure o demad parameter BÖRCSÖK J., SCHAEFER S. Departmet of Computer Architecture ad System Programmig Uiversity Kassel, Wilhelmshöher Allee

More information

Evaluation of Different Fitness Functions for the Evolutionary Testing of an Autonomous Parking System

Evaluation of Different Fitness Functions for the Evolutionary Testing of an Autonomous Parking System Evaluatio of Differet Fitess Fuctios for the Evolutioary Testig of a Autoomous Parkig System Joachim Wegeer 1 ad Oliver Bühler 2 1 DaimlerChrysler AG, Research ad Techology, Alt-Moabit 96 a, D-10559 Berli,

More information

Keywords Software Architecture, Object-oriented metrics, Reliability, Reusability, Coupling evaluator, Cohesion, efficiency

Keywords Software Architecture, Object-oriented metrics, Reliability, Reusability, Coupling evaluator, Cohesion, efficiency Volume 3, Issue 9, September 2013 ISSN: 2277 128X Iteratioal Joural of Advaced Research i Computer Sciece ad Software Egieerig Research Paper Available olie at: www.ijarcsse.com Couplig Evaluator to Ehace

More information

EMPIRICAL ANALYSIS OF FAULT PREDICATION TECHNIQUES FOR IMPROVING SOFTWARE PROCESS CONTROL

EMPIRICAL ANALYSIS OF FAULT PREDICATION TECHNIQUES FOR IMPROVING SOFTWARE PROCESS CONTROL Iteratioal Joural of Iformatio Techology ad Kowledge Maagemet July-December 2012, Volume 5, No. 2, pp. 371-375 EMPIRICAL ANALYSIS OF FAULT PREDICATION TECHNIQUES FOR IMPROVING SOFTWARE PROCESS CONTROL

More information

Evaluation of Distributed and Replicated HLR for Location Management in PCS Network

Evaluation of Distributed and Replicated HLR for Location Management in PCS Network JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 9, 85-0 (2003) Evaluatio of Distributed ad Replicated HLR for Locatio Maagemet i PCS Network Departmet of Computer Sciece ad Iformatio Egieerig Natioal Chiao

More information

HADOOP: A NEW APPROACH FOR DOCUMENT CLUSTERING

HADOOP: A NEW APPROACH FOR DOCUMENT CLUSTERING Y.K. Patil* Iteratioal Joural of Advaced Research i ISSN: 2278-6244 IT ad Egieerig Impact Factor: 4.54 HADOOP: A NEW APPROACH FOR DOCUMENT CLUSTERING Prof. V.S. Nadedkar** Abstract: Documet clusterig is

More information

OCR Statistics 1. Working with data. Section 3: Measures of spread

OCR Statistics 1. Working with data. Section 3: Measures of spread Notes ad Eamples OCR Statistics 1 Workig with data Sectio 3: Measures of spread Just as there are several differet measures of cetral tedec (averages), there are a variet of statistical measures of spread.

More information

1 Enterprise Modeler

1 Enterprise Modeler 1 Eterprise Modeler Itroductio I BaaERP, a Busiess Cotrol Model ad a Eterprise Structure Model for multi-site cofiguratios are itroduced. Eterprise Structure Model Busiess Cotrol Models Busiess Fuctio

More information

ON THE QUALITY OF AUTOMATIC RELATIVE ORIENTATION PROCEDURES

ON THE QUALITY OF AUTOMATIC RELATIVE ORIENTATION PROCEDURES ON THE QUALITY OF AUTOMATIC RELATIVE ORIENTATION PROCEDURES Thomas Läbe, Timo Dickscheid ad Wolfgag Förster Istitute of Geodesy ad Geoiformatio, Departmet of Photogrammetry, Uiversity of Bo laebe@ipb.ui-bo.de,

More information

Research on Identification Model of Financial Fraud of Listed Company Based on Data Mining Technology

Research on Identification Model of Financial Fraud of Listed Company Based on Data Mining Technology 208 2d Iteratioal Coferece o Systems, Computig, ad Applicatios (SYSTCA 208) Research o Idetificatio Model of Fiacial Fraud of Listed Compay Based o Data Miig Techology Jiaqi Hu, Xiao Che School of Busiess,

More information

Analysis of Documents Clustering Using Sampled Agglomerative Technique

Analysis of Documents Clustering Using Sampled Agglomerative Technique Aalysis of Documets Clusterig Usig Sampled Agglomerative Techique Omar H. Karam, Ahmed M. Hamad, ad Sheri M. Moussa Abstract I this paper a clusterig algorithm for documets is proposed that adapts a samplig-based

More information

Image based Cats and Possums Identification for Intelligent Trapping Systems

Image based Cats and Possums Identification for Intelligent Trapping Systems Volume 159 No, February 017 Image based Cats ad Possums Idetificatio for Itelliget Trappig Systems T. A. S. Achala Perera School of Egieerig Aucklad Uiversity of Techology New Zealad Joh Collis School

More information

A QoS Provisioning mechanism of Real-time Wireless USB Transfers for Smart HDTV Multimedia Services

A QoS Provisioning mechanism of Real-time Wireless USB Transfers for Smart HDTV Multimedia Services A QoS Provisioig mechaism of Real-time Wireless USB Trasfers for Smart HDTV Multimedia Services Ji-Woo im 1, yeog Hur 2, Jog-Geu Jeog 3, Dog Hoo Lee 4, Moo Sog Yeu 5, Yeowoo Lee 6 ad Seog Ro Lee 7 1 Istitute

More information