Face Anti-spoofing based on Deep Stack Generalization Networks

Size: px
Start display at page:

Download "Face Anti-spoofing based on Deep Stack Generalization Networks"

Transcription

1 Xi Nig,, Weiju Li,, Meili Wei, Liju Su, ad Xiaoli Dog, Istitute of Semicoductors, Chiese Academy of Scieces, 00083, Beijig, Chia Cogitive Computig Techology Wave Joit Lab, 00083, Beijig, Chia Keywords: Abstract: Face Ati-spoofig, Covolutioal Neural Networks, Stacked Geeralized Approach, Itra-class Variatios. Thaks for the recet developmet of Covolutioal Neural Networks (CNNs), the performace of face ati-spoofig methods has bee improved by extractig more distiguishig features betwee geuie ad fake faces tha the had-crafted texture features. As kow, the way of fraud is diverse, thus the fake class has large itra-class variatios, so traiig as a biary classificatio problem is hard to lear the distiguishig features. I this work, our cotributio is a ovel model fusio approach for face atispoofig, which ca reduce the itra-class variatios. Accordig to the type of fraud, we firstly trai differet models for face ati-spoofig problem by CNN, thus the itra-class variatios of fake class has reduced durig traiig each model. Distiguishig features ca be leared more easily. The the stacked geeralized method is used for combiig the lower models to achieve better predictive accuracy. For perfectig the geeralized accuracy, the stacked geeralized approach chages the weight of each model's predictio, so that the model after fusio ca predict precisely whether the face image is fake or geuie. Meawhile, the experimetal results idicate our method ca obtai excellet results compared to the stateof-the-art methods. INTRODUCTION Face recogitio has achieved great success durig the past decades, ad has bee widely applied i access cotrol system, logi system ad so o. However, a high security requiremet for face autheticatio is urget, because oly a photo, video replay, or 3D-mask of valid user ca easily spoof a face recogitio system to access secure iformatio illegally. With the popularity of electroic devices, people ca easily get photos ad videos belogig to others through the etwork, which caused a lot of face recogitio system for security risks. Therefore, ati-spoofig problem for face biometric system has gaied great attetio to the scholars ad compaies. The fragility of face recogitio systems to face spoof attacks has motivated a umber of studies o face ati-spoofig, such as LBP (Maatta et al., 0), HOG (Komulaie et al., 03), LBP-TOP (Pereira, 0), Image quality assessmet (We et al., 05), CNN (Yag et al., 04), etc. However, published studies are limited i their scope, because these methods are more to trai a commo model to prevet attack from photo, video or 3D-mask. Because of the diversity of fraud, a geeral model caot be leared i such complexities, may ot work whe facig specific fraud. Besides, these methods regard face atispoofig as a biary classificatio problem, that is to say the all kids of fake face is oe class, ad the real face is aother class. As we all kow, the fake face ca be a photo, a video replay, or a 3D-mask. Due to differet ways of fraud, the fake class has large itra-class variatios. These existig classifiers i idetifyig sample work idividually, as is well kow, whe makig critical decisios, wise people ofte take ito accout the opiios from several experts rather tha oly oe. I this paper, we do ot address 3D-mask attacks, which are more costly to lauch, we focus o photo ad replayed video attacks. We lear a CNN model for each type of fraud, the the stacked geeralizatio method will be use to itegrate the two models, which will decide whether the iput face image is real or ot. Stacked geeralizatio is a geeral method of usig a highlevel model to combie lower-level models to achieve greater predictive accuracy. The remaider of the paper is orgaized as follows: Sectio briefly reviews the relevat work o face ati-spoofig. Sectio 3 presets our 37 Nig, X., Li, W., Wei, M., Su, L. ad Dog, X.. DOI: 0.50/ I Proceedigs of the 7th Iteratioal Coferece o Patter Recogitio Applicatios ad Methods (ICPRAM 08), pages ISBN: Copyright 08 by SCITEPRESS Sciece ad Techology Publicatios, Lda. All rights reserved

2 ICPRAM 08-7th Iteratioal Coferece o Patter Recogitio Applicatios ad Methods approach. The experimetal setup ad results are discussed i Sectio 4. Fially, i Sectio 5 we summarize this work highlightig its mai cotributios. RELATED WORK Because of the diversity of spoofig attacks, existig traditioal face ati-spoofig approaches ca be maily categorized ito four categories: (i) motio based methods, (ii) texture based methods, (iii) method based o image quality aalysis, ad (iv) methods based o other cues. The motio based methods was desiged primarily to couter prited photo attacks. Eye blikig (Pa et al., 007) or lip movemet (Kollreider et al., 007) are used for face ati-spoofig. Give that motio is a relative feature across video frames, these methods are expected to have better geeralizatio ability tha the texture based methods. However, motio based methods eed a relatively log time to accumulate stable vitality features for face spoof detectio. The texture based methods iclude LBP (Maatta et al., 0), HOG (Komulaie et al., 03), etc. Pereira et al. (0) used LBP-TOP to extract spatial ad time domai features from three orthogoal plaes. Ulike motio based methods, texture based methods eed oly a sigle image to detect a spoof. However, the geeralizatio ability of may texture based methods has bee foud to be poor. A recet work (Galbally et al., 04) proposed a biometric liveess detectio method for iris, figerprit ad face images usig 5 image quality measures, icludig full-referece measures ad 4 oreferece measures. Differet from traditioal methods, CNNs ca extract distiguishig ed-to-ed features directly from raw data, ad has bee proved efficiet i may other visio fields. Yag et al., (04) extract features by CNN, the feedig them to a SVM classifier. Xu et al., (06) proposed LSTM-CNN architecture to lear the temporal structure from videos. Those works cosider the face ati-spoofig as a biary classificatio problem, all real face is oe class, ad the other is all kids of fake face. Because of the variety of fake face, photo attacks ad video attacks will be differet o the texture, reflect illumiatio, resolutio, etc., thus the large itravariace will icrease the difficulty of classificatio. Each model is heterogeeous ad has strog classificatio ability, therefore, the itegratio model with stacked geeralizatio method will make full use of the advatages of differet models, complemet each other, thus will achieve better predictive accuracy. 3 PROPOSED METHOD For reducig the itra-class variatios, we trai differet models accordig to the type of fraud face. Each model ca lear the deep ad distiguishig feature for classifyig fake or real. The stacked geeralizatio method takes full advatage of each model's predictio ad chage the weights of each predictio. The a wiser decisio would be made for maximizig geeralize accuracy. With the stacked geeralized method, the traiig difficulty of atispoofig problem is decreased tha traiig a geeral model. Besides, the model ca coverge more easily. 3. Stacked Geeralizatio Stacked geeralizatio (Wolpert, 99; Tig ad Witte, 997) is a geeral method of usig a highlevel model to combie lower-level models to achieve greater predictive accuracy. It's a scheme for miimizig the geeralizatio error rate of oe or more classifiers, ad works by reducig the biases of the classifiers with respect to a provided learig set. Whe fusig the multiple classifiers, stacked geeralizatio exploited a strategy more sophisticated for combiig the idividual classifiers. Stacked geeralizatio tries to lear which classifiers are reliable oes, ad use a higherlevel learig algorithm, the so-called "metaclassifier", to discover the best way of how to combie the outputs of the base classifiers. As show i Figure, there are two kids of classifiers: several base classifiers (leavel-0 classifiers) ad oe meta-classifier (level- classifier). The output class probabilities geerated by level-0 models are used to form level- data. The a multivariable liear regressio model (MLR) is adapted for classificatio tasks for level- classifier. 3.. Level-0 Geeralizers As show before (Krizhevsky et al., 0; Simoya ad Zisserma, 04; He et al., 06), deeper ad better pre-traiig etworks lead to better performace. ResNet (He et al., 06) wo the st place o the ILSVRC 05 classificatio task. The depth of represetatios is of cetral importace for may visual recogitio tasks, especially i face 38

3 ati-spoofig. I the task of face ati-spoofig, the itra-class variatios are large, ad are maily caused by the appearace of differet people, the differet ways of fraud, ad the differet resolutios of faces images captured by differet camera, such as the photo attacks ca be prited o differet types of paper, the video attacks ca be played o differet electroic equipmet ad so o. Therefore, the deeper etworks ca extract more useful ad distiguishig features. I this paper, we use a ResNet50 (He et al., 06) model as the level-0 geeralizers. Give two traiig data sets Q y, x,,..., N ad P y, x,,..., N,where y is the class value ad x represets the attribute values of the ss istace, Q defies the set of real face ad video attack face, ad P defies the set of real face ad photo attack face. The specific type of fraud is icluded i each dataset, causig the small itraclass variatios ad large iter-class variatios. So the traiig difficulty ca be decreased. The radomly split the data sets ito two parts Q, P ad Q, P. Defie P y, x,,..., N M Q y, x,,..., M P y, x,,..., N M Q y, x,,..., N M, ad, to be the traiig ad validatio sets. The Q ad P datasets are used for traiig level-0 data, ad the Q ad P datasets are used for formig the level- data. Give ResNet50 called level-0 geeralizers, traiig o the data i the traiig set Q ad P to iduce a model, for k,..., K, which are called level-0 models. The level- data assembled from the outputs of the K models is ' Q y, z,..., z I,..., zk,..., zki,..., zk,..., zki,,..., M, ' P y, z,..., z,..., z,..., z,..., z,..., z,,..., M, where zki Pki x 0 models, ad I k ki K KI is the predictio from the level- Pki x deote the probability of the ith output class, ad i the ResNet50, Pki x is the class probability of the Softmax layer's output, ad I Pki x. i Figure : The illustratio of stacked geeralizatio. 3.. Level- Geeralizers After obtaiig the level- data, we use MLR (Breima, 996) to derive from this data a model i this work. MLR which Breima (Breima, 996) used i regressio settigs, is a adaptatio of a least-square liear regressio algorithm. If the classificatio problem is with real-valued attributes, it ca be trasformed ito a multi-respose regressio problem. I the face ati-spoofig task, they are two classes, they ca be coverted ito two separate regressio problems, where the problem for class has istaces with resposes equal to whe they have class ad 0 otherwise. The liear regressio for class is simply: K k I i ki ki LR z P z () We solve this problem to choose the liear regressio coefficiets to miimize: ki y ki Pki z () j ' y, z Q k i where y is equal to whe the istace's label is correspodig with the class,s ad 0 otherwise. These regressio coefficiets ca be the weight of each level-0 model's predictio probability. I the test phase, after the fusio, the probability of each sample predictig fake or real is readjusted, so that accordig to the probability, the sample is classifyig correctly. I the experimetal stage, we compare the accuracy with the AND rule. As show i Table, our method is better tha the AND rule with a large margi. Accordig the AND rule, if both the two model's predictio is real, the the fial predictio is real, otherwise is fake. We believe the 39

4 ICPRAM 08-7th Iteratioal Coferece o Patter Recogitio Applicatios ad Methods reaso is that our method lears which model is more reliable ad which ot. Because each level-0 model is heterogeeous, for each specific sample, the classificatio accuracy rates of these models are quite differet. Therefore, the weights ca adjust the predictios so that the more wisdom decisio ca be give. Whe classifyig a ew istace x, compute LR x for the two classes, ad assig the istace to class real which has the greatest value. That is to LR x LR x, the we believe the say, if real fake sample is a real face, otherwise fake face. Figure : Examples of real access ad spoofig attempts i the CASIA-FASD database. The first row is low-quality, the secod row is middle-quality ad the last row is highquality. The first colum is geuie, ad the others form left to right are warp-photo, cut-photo ad video fraud. Figure 4: Examples of real access ad spoofig attempts i our database. The five rows are captured by differet cameras, from top to dow are Phoe Low Resolutio frot camera, Pad frot camera, Phoe Normal Resolutio frot camera, USB Low Resolutio ad USB High Resolutio. The four colums are differet type of fraud, from left to right are warp A4 photo fraud, cut eye A4 photo, warp copper photo ad video replay fraud. 4 EXPERIMENTS Our proposed method is successful o the face atispoofig data sets. We evaluated it i this sectio o two differet datasets, icludig CASIA (Zhag et al., 0), ad Replay-attack (Chigovska et al., 0). 4. Implemetatio Details Figure 3: Examples of real access ad spoofig attempts i the REPLAY-ATTACK database. The four rows from top to dow are real video, prit fraud video, mobile fraud video ad high-defiitio fraud video. Ad the first colum is adverse eviromet ad the secod colum is cotrolled. I this work, durig the traiig, we first separate the video attacks ad photo attacks ito differet sets. For each video, we selected oe frame every three frames, thus formig the traiig, validatio ad test sample sets. For each frame, the face ca be detected by usig Viola-Joes algorithm. To provide precise face locatios, we implemet the face aligmet algorithm proposed i Su et al., (04) 30

5 after a commo Viola-Joes face detectio. After obtaiig the ladmarks, their boudig box is regarded as the fial face locatio. Accordig to Yag et al. (04), beyod the covetioal face regio, the backgrouds are helpful for the classificatio as well. Therefore, we elarge the origial oes with re-scalig ratio.8. Fially, all iput images are resized to 4*4. For the CNN, we use Caffe toolbox ad adopt a commoly used structure ResNet50, which was ever used i He et al. (06). I the traiig of the ResNet50 for video attack face ad photo attack face, the learig rate is 0.000; decay rate is 0.00; ad the mometum is 0.9. Before fed ito the ResNet50, the data are first cetralized by the mea of traiig data. 4. Datasets 4.. CASIA-FASD Database The database collects 600 short videos from 50 cliets (Zhag et al., 0). For each subject, there differet quality videos are captured by differet resolutio camera. For each camera, oe real video ad correspodig three differet kids of attacks were recorded. The three kids of attacks are warped photo attack, cut photo attack ad electroic scree attack. Therefore, each subject has sequeces (3 geuie ad 9 fake oes). Three differet imagig quality coditios were recorded usig a imagig device of () High-quality, () Middle-quality, ad (3) Low-quality. Example frames from geuie ad fake videos are show i Figure. For evaluatio, the total set of videos is divided ito two ooverlappig subsets for traiig ad testig. 4.. REPLAY-ATTACK Database This database cotais 00 short videos from 50 subjects (Chigovska et al., 0), icludig both real accesses ad face spoofig attacks. Each perso i the database was recorded the videos i two illumiatio coditios: cotrolled ad adverse. A high resolutio pictures ad videos were take for each subject uder the same coditio. There are three attacks: prit attacks, mobile attacks ad highdefiitio attacks. Ad the videos were divided ito had based attacks ad fixed based attacks. Example frames from geuie ad fake videos are show i Figure 3.For evaluatio, the total set of videos is divided ito three o-overlappig subsets for traiig, developmet ad testig Our Database The database cotais 500 short videos from 00 subjects, icludig both real accesses ad face spoofig attacks. For each subject, there differet quality videos are captured by differet resolutio camera. For each camera, oe real video ad correspodig three differet kids of attacks were recorded. The three kids of attacks are warped photo attack, cut photo attack ad electroic scree attack. Example frames from geuie ad fake videos are show i Figure 4. Therefore, each subject has 5 sequeces ( geuie ad 4 fake oes). For evaluatio, the total set of videos is divided ito two o-overlappig subsets for traiig. 4.3 Experimetal Results The performace of the proposed stack geeralized based face ati-spoofig approach was evaluated o the three databases. All these results are give i Table. For a performace compariso, the results of the state-of-the-art coutermeasures ad the baselie algorithms i databases to face spoofig attacks are listed i Table. O the CASIA-FASD database, best performace i previous work was achieved by the LBPs form three orthogoal plaes (LBP-TOP) method, explorig the spatial ad temporal LBP distributios simultaeously. Our method achieved a EER of 3.4%, which is better tha the LBP-TOP method. O the REPLAY- ATTACK, our method achieved a EER of 0.3% ad HTER of 0.63% respectively, both of which are superior to the others. Besides, the performace of our method was compared with the AND rule method o the REPLAY-ATTACK, CASIA-FASD ad our database. The results are listed i Table.From the results, the proposed stacked geeralized method achieved a huge performace improvemet i liveess detectio compared with the AND rule method. O the REPLAY-ATTACK, CASIA-FASD ad our database, our method achieved a huge performace improvemet i face ati-spoofig problem. These results illustrate the effectiveess of the proposed stacked geeralized face liveess detectio approach. 3

6 ICPRAM 08-7th Iteratioal Coferece o Patter Recogitio Applicatios ad Methods Table : Compariso betwee the proposed coutermeasure ad state-of-the-art methods based o the REPLAY-ATTACK, CASIA-FASD database. Approach IQA based (Galbally ad Marcel, 04) Motio (Pereira et al., 03 ) LBP + SVM (Yag et al., 03) LBP-TOP + SVM (Pereira et al., 0) SBIQF+NN (Feg et al., 06 ) YCbCr + HSV + LBP (Boulkeafet et al., 05) LSTM (Xu et al., 06) Replayattackdev(EER)% Replayattacktest(HTER)% CASIA- FASD-test- EER (%) Our Method Table : Compariso betwee the proposed coutermeasure ad the AND rule method based o the REPLAY-ATTACK, CASIA-FASD database. Approach Replay-attackdev(EER)% Replay-attacktest(HTER)% CASIA-FASDtest(EER)% AND Rule Our Method CONCLUSIONS With the rapid developmet of face ati-spoofig techiques, the threats of spoofig attacks will also icrease i the diversity. It's hard to lear a model to detect all types of fraud. Hece, the itegratio of several coutermeasures is a promisig approach. The proposed method is a way of combiatio. Ad our method ca be easily combied with other algorithms, as log as these algorithms are helpful for liveess detectio. I future work, other advace eural etworks will be ivestigated to improve face ati-spoofig performace, such as the log shortterm memory (LSTM) etwork, which may be more effective i learig face liveess features. ACKNOWLEDGEMENTS This work was supported by the Natioal Nature Sciece Foudatio of Chia (No ), ad the Chia Scholarship Coucil (Grat No ). REFERENCES Maatta J, Hadid A, Pietikaie M., 0. Face spoofig detectio from sigle images usig micro-texture aalysis[c]// Iteratioal Joit Coferece o Biometrics. IEEE. Komulaie J, Hadid A, Pietikaie M., 03. Cotext based face ati-spoofig[c]// IEEE Sixth Iteratioal Coferece o Biometrics: Theory, Applicatios ad Systems. IEEE. Pereira T D F, Ajos A, Martio J M D, et al., 0. LBP TOP Based Coutermeasure agaist Face Spoofig Attacks [J]. We D, Ha H, Jai A K., 05. Face Spoof Detectio with Image Distortio Aalysis [J]. IEEE Trasactios o Iformatio Foresics & Security. Yag J, Lei Z, Li S Z., 04. Lear Covolutioal Neural Network for Face Ati-Spoofig [J]. Computer Sciece. Pa G, Su L, Wu Z, et al., 007. Eyeblik-based Ati- Spoofig i Face Recogitio from a Geeric Webcamera[C]// IEEE, Iteratioal Coferece o Computer Visio. Kollreider K, Frothaler H, Faraj M I, et al., 007. Real- Time Face Detectio ad Motio Aalysis with Applicatio i Liveess Assessmet [J]. Iformatio Forssssesics & Security IEEE Trasactios o. Galbally J, Marcel S, Fierrez J., 04. Image Quality Assessmet for Fake Biometric Detectio: Applicatio to Iris, Figerprit, ad Face Recogitio [J]. IEEE Trasactios o Image Processig a Publicatio of the IEEE Sigal Processig Society. Xu Z, Li S, Deg W., 06 Learig temporal features usig LSTM-CNN architecture for face atispoofig[c]// Patter Recogitio. IEEE. Wolpert D H., 99. Origial Cotributio: Stacked geeralizatio [J]. Neural Netw. Tig K M, Witte I H., 997. Stacked geeralizatio: whe does it work? [C]// Fifteeth Iteratioal Joit Coferece o Artificial Itelligece. Morga Kaufma Publishers Ic. Krizhevsky A, Sutskever I, Hito G E., 0. ImageNet classificatio with deep covolutioal eural etworks[c]// Iteratioal Coferece o Neural Iformatio Processig Systems. Curra Associates Ic. Simoya K, Zisserma A., 04 Very Deep Covolutioal Networks for Large-Scale Image Recogitio [J]. Computer Sciece. He K, Zhag X, Re S, et al., 06. Deep Residual Learig for Image Recogitio[C]// Computer Visio ad Patter Recogitio. IEEE. Breima L., 996. Stacked regressios [J]. Machie Learig. 3

7 Chigovska I, Ajos A, Marcel S., 0. O the effectiveess of local biary patters i face atispoofig[c]// Biometrics Special Iterest Group. IEEE. Zhag Z, Ya J, Liu S, et al., 0. A face at spoofig database with diverse attacks[c]// Iapr Iteratioal Coferece o Biometrics. Galbally J, Marcel S., 04. Face Ati-spoofig Based o Geeral Image Quality Assessmet[C]// Iteratioal Coferece o Patter Recogitio. IEEE Computer Society. Pereira T D F, Ajos A, Martio J M D, et al., 03. Ca face ati-spoofig coutermeasures work i a real world sceario? [C]// Iteratioal Coferece o Biometrics. IEEE. Yag J, Lei Z, Liao S, et al., 03. Face liveess detectio with compoet depedet descriptor[c]// Iteratioal Coferece o Biometrics. IEEE. Pereira T D F, Ajos A, Martio J M D, et al., 0. LBP TOP, Based Coutermeasure agaist Face Spoofig Attacks[C]// Iteratioal Coferece o Computer Visio. Feg L, Po L M, Li Y, et al., 06. Itegratio of image quality ad motio cues for face ati-spoofig [J]. Joural of Visual Commuicatio & Image Represetatio. Boulkeafet, Zielabidie, Jukka Komulaie, ad Abdeour Hadid., 05 "Face ati-spoofig based o color texture aalysis." Image Processig (ICIP), 05 IEEE Iteratioal Coferece o. IEEE. Xu Z, Li S, Deg W., 06. Learig temporal features usig LSTM-CNN architecture for face atispoofig[c]// Patter Recogitio. IEEE. Su J, We F, Wei Y, et al., 04. Face aligmet by Explicit Shape Regressio [J]. Iteratioal Joural of Computer Visio. 33

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

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

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

An Improved Shuffled Frog-Leaping Algorithm for Knapsack Problem

An Improved Shuffled Frog-Leaping Algorithm for Knapsack Problem A Improved Shuffled Frog-Leapig Algorithm for Kapsack Problem Zhoufag Li, Ya Zhou, ad Peg Cheg School of Iformatio Sciece ad Egieerig Hea Uiversity of Techology ZhegZhou, Chia lzhf1978@126.com Abstract.

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

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 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

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

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

Investigating methods for improving Bagged k-nn classifiers

Investigating methods for improving Bagged k-nn classifiers Ivestigatig methods for improvig Bagged k-nn classifiers Fuad M. Alkoot Telecommuicatio & Navigatio Istitute, P.A.A.E.T. P.O.Box 4575, Alsalmia, 22046 Kuwait Abstract- We experimet with baggig knn classifiers

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

Designing a learning system

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

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

Deep Joint Discriminative Learning for Vehicle Re-identification and Retrieval

Deep Joint Discriminative Learning for Vehicle Re-identification and Retrieval Deep Joit Discrimiative Learig for Vehicle Re-idetificatio ad Retrieval Yuqi Li, Yaghao Li, Hogfei Ya, Jiayig Liu Pekig Uiversity 2017.09.18 Deep Joit Discrimiative Learig for Vehicle Re-idetificatio ad

More information

An Anomaly Detection Method Based On Deep Learning

An Anomaly Detection Method Based On Deep Learning Sed Orders for Reprits to reprits@bethamsciece.ae 734 The Ope Automatio ad Cotrol Systems Joural, 05, 7, 734-739 A Aomaly Detectio Method Based O Deep Learig Ope Access Hog-li Deg *, Tao yag ad Jiag-i

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

Fragments based tracking with adaptive multi-cue integration

Fragments based tracking with adaptive multi-cue integration Fragmets based tracig with adaptive multi-cue itegratio Abstract Lichua Gu *, Jiaxiao Liu, Chegi Wag School of Computer ad Iformatio, Ahui Agriculture Uiversity, Hefei Ahui, No.130 West Chagiag Road Chia,

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

New Fuzzy Color Clustering Algorithm Based on hsl Similarity

New Fuzzy Color Clustering Algorithm Based on hsl Similarity IFSA-EUSFLAT 009 New Fuzzy Color Clusterig Algorithm Based o hsl Similarity Vasile Ptracu Departmet of Iformatics Techology Tarom Compay Bucharest Romaia Email: patrascu.v@gmail.com Abstract I this paper

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

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

A new algorithm to build feed forward neural networks.

A new algorithm to build feed forward neural networks. A ew algorithm to build feed forward eural etworks. Amit Thombre Cetre of Excellece, Software Techologies ad Kowledge Maagemet, Tech Mahidra, Pue, Idia Abstract The paper presets a ew algorithm to build

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

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

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

The identification of key quality characteristics based on FAHP

The identification of key quality characteristics based on FAHP Iteratioal Joural of Research i Egieerig ad Sciece (IJRES ISSN (Olie: 2320-9364, ISSN (Prit: 2320-9356 Volume 3 Issue 6 ǁ Jue 2015 ǁ PP.01-07 The idetificatio of ey quality characteristics based o FAHP

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

An improved support vector machine based on particle swarm optimization in laser ultrasonic defect detection

An improved support vector machine based on particle swarm optimization in laser ultrasonic defect detection A improved support vector machie based o particle swarm optimizatio i laser ultrasoic defect detectio School of Sciece, North Uiversity of Chia, aiyua, Shaxi 35, Chia xut98@63.com,hhp9@63.com,xywag@6.com,43497@qq.com

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

Chapter 3 Classification of FFT Processor Algorithms

Chapter 3 Classification of FFT Processor Algorithms Chapter Classificatio of FFT Processor Algorithms The computatioal complexity of the Discrete Fourier trasform (DFT) is very high. It requires () 2 complex multiplicatios ad () complex additios [5]. As

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

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

Image Restoration from Patch-based Compressed Sensing Measurement

Image Restoration from Patch-based Compressed Sensing Measurement Image Restoratio from Patch-based Compressed Sesig Measuremet Guagtao Nie 1, Yig Fu 1, Yiqiag Zheg 2, Hua Huag 1 1 Beijig Istitute of Techology, 2 Natioal Istitute of Iformatics {lightbillow,fuyig,huahuag}@bit.edu.c,

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

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

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

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

Mapping Publishing and Mapping Adaptation in the Middleware of Railway Information Grid System

Mapping Publishing and Mapping Adaptation in the Middleware of Railway Information Grid System Mappig Publishig ad Mappig Adaptatio i the Middleware of Railway Iformatio Grid ystem You Gamei, Liao Huamig, u Yuzhog Istitute of Computig Techology, Chiese Academy of cieces, Beijig 00080 gameiu@ict.ac.c

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

GPUMP: a Multiple-Precision Integer Library for GPUs

GPUMP: a Multiple-Precision Integer Library for GPUs GPUMP: a Multiple-Precisio Iteger Library for GPUs Kaiyog Zhao ad Xiaowe Chu Departmet of Computer Sciece, Hog Kog Baptist Uiversity Hog Kog, P. R. Chia Email: {kyzhao, chxw}@comp.hkbu.edu.hk Abstract

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

Neural Networks A Model of Boolean Functions

Neural Networks A Model of Boolean Functions Neural Networks A Model of Boolea Fuctios Berd Steibach, Roma Kohut Freiberg Uiversity of Miig ad Techology Istitute of Computer Sciece D-09596 Freiberg, Germay e-mails: steib@iformatik.tu-freiberg.de

More information

APPLICATION NOTE. Automated Gain Flattening. 1. Experimental Setup. Scope and Overview

APPLICATION NOTE. Automated Gain Flattening. 1. Experimental Setup. Scope and Overview APPLICATION NOTE Automated Gai Flatteig Scope ad Overview A flat optical power spectrum is essetial for optical telecommuicatio sigals. This stems from a eed to balace the chael powers across large distaces.

More information

Facial Expression Recognition Based on Histogram Sequence of Local Gabor Binary Patterns

Facial Expression Recognition Based on Histogram Sequence of Local Gabor Binary Patterns Facial Expressio Recogitio Based o Histogram Sequece of Local Gabor Biary Patters Xighua Su, Hogxia Xu, Chuxia Zhao, ad Jigyu Yag School of Computer Sciece ad Techology Najig Uiversity of Sciece ad Techology

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

Our second algorithm. Comp 135 Machine Learning Computer Science Tufts University. Decision Trees. Decision Trees. Decision Trees.

Our second algorithm. Comp 135 Machine Learning Computer Science Tufts University. Decision Trees. Decision Trees. Decision Trees. Comp 135 Machie Learig Computer Sciece Tufts Uiversity Fall 2017 Roi Khardo Some of these slides were adapted from previous slides by Carla Brodley Our secod algorithm Let s look at a simple dataset for

More information

Software Fault Prediction of Unlabeled Program Modules

Software Fault Prediction of Unlabeled Program Modules Software Fault Predictio of Ulabeled Program Modules C. Catal, U. Sevim, ad B. Diri, Member, IAENG Abstract Software metrics ad fault data belogig to a previous software versio are used to build the software

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

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

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

performance to the performance they can experience when they use the services from a xed location.

performance to the performance they can experience when they use the services from a xed location. I the Proceedigs of The First Aual Iteratioal Coferece o Mobile Computig ad Networkig (MobiCom 9) November -, 99, Berkeley, Califoria USA Performace Compariso of Mobile Support Strategies Rieko Kadobayashi

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

What are Information Systems?

What are Information Systems? Iformatio Systems Cocepts What are Iformatio Systems? Roma Kotchakov Birkbeck, Uiversity of Lodo Based o Chapter 1 of Beett, McRobb ad Farmer: Object Orieted Systems Aalysis ad Desig Usig UML, (4th Editio),

More information

Handwriting Stroke Extraction Using a New XYTC Transform

Handwriting Stroke Extraction Using a New XYTC Transform Hadwritig Stroke Etractio Usig a New XYTC Trasform Gilles F. Houle 1, Kateria Bliova 1 ad M. Shridhar 1 Computer Scieces Corporatio Uiversity Michiga-Dearbor Abstract: The fudametal represetatio of hadwritig

More information

Performance Comparisons of PSO based Clustering

Performance Comparisons of PSO based Clustering Performace Comparisos of PSO based Clusterig Suresh Chadra Satapathy, 2 Guaidhi Pradha, 3 Sabyasachi Pattai, 4 JVR Murthy, 5 PVGD Prasad Reddy Ail Neeruoda Istitute of Techology ad Scieces, Sagivalas,Vishaapatam

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

BOOLEAN MATHEMATICS: GENERAL THEORY

BOOLEAN MATHEMATICS: GENERAL THEORY CHAPTER 3 BOOLEAN MATHEMATICS: GENERAL THEORY 3.1 ISOMORPHIC PROPERTIES The ame Boolea Arithmetic was chose because it was discovered that literal Boolea Algebra could have a isomorphic umerical aspect.

More information

BASED ON ITERATIVE ERROR-CORRECTION

BASED ON ITERATIVE ERROR-CORRECTION A COHPARISO OF CRYPTAALYTIC PRICIPLES BASED O ITERATIVE ERROR-CORRECTIO Miodrag J. MihaljeviC ad Jova Dj. GoliC Istitute of Applied Mathematics ad Electroics. Belgrade School of Electrical Egieerig. Uiversity

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

Model Enhancement in Data Mining: Calibration, ROC Analysis, Model Combination and Mimetic Models

Model Enhancement in Data Mining: Calibration, ROC Analysis, Model Combination and Mimetic Models Model Ehacemet i Miig: Calibratio, ROC Aalysis, Model Combiatio ad Mimetic Models José Herádez-Orallo Dpto. de Sistemas Iformáticos y Computació, Uiversidad Politécica de Valecia, jorallo@dsic.upv.es Rome,

More information

Multi-Agent Decision Tree Learning from Distributed Autonomous Data. Sources. D. Caragea, A. Silvescu, and V. Honavar

Multi-Agent Decision Tree Learning from Distributed Autonomous Data. Sources. D. Caragea, A. Silvescu, and V. Honavar Multi-Aget Decisio Tree Learig from Distributed Autoomous Data Sources D. Caragea, A. Silvescu, ad V. Hoavar Iowa State Uiversity Computer Sciece Departmet Artificial Itelligece Research Group Ames, IA

More information

Eigenimages. Digital Image Processing: Bernd Girod, Stanford University -- Eigenimages 1

Eigenimages. Digital Image Processing: Bernd Girod, Stanford University -- Eigenimages 1 Eigeimages Uitary trasforms Karhue-Loève trasform ad eigeimages Sirovich ad Kirby method Eigefaces for geder recogitio Fisher liear discrimat aalysis Fisherimages ad varyig illumiatio Fisherfaces vs. eigefaces

More information

We are IntechOpen, the first native scientific publisher of Open Access books. International authors and editors. Our authors are among the TOP 1%

We are IntechOpen, the first native scientific publisher of Open Access books. International authors and editors. Our authors are among the TOP 1% We are ItechOpe, the first ative scietific publisher of Ope Access books 3,350 108,000 1.7 M Ope access books available Iteratioal authors ad editors Dowloads Our authors are amog the 151 Coutries delivered

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

Redundancy Allocation for Series Parallel Systems with Multiple Constraints and Sensitivity Analysis

Redundancy Allocation for Series Parallel Systems with Multiple Constraints and Sensitivity Analysis IOSR Joural of Egieerig Redudacy Allocatio for Series Parallel Systems with Multiple Costraits ad Sesitivity Aalysis S. V. Suresh Babu, D.Maheswar 2, G. Ragaath 3 Y.Viaya Kumar d G.Sakaraiah e (Mechaical

More information

Intrusion Detection using Fuzzy Clustering and Artificial Neural Network

Intrusion Detection using Fuzzy Clustering and Artificial Neural Network Itrusio Detectio usig Fuzzy Clusterig ad Artificial Neural Network Shraddha Suraa Research Scholar, Departmet of Computer Egieerig, Vishwakarma Istitute of Techology, Pue Idia shraddha.suraa@gmail.com

More information

Introduction. Nature-Inspired Computing. Terminology. Problem Types. Constraint Satisfaction Problems - CSP. Free Optimization Problem - FOP

Introduction. Nature-Inspired Computing. Terminology. Problem Types. Constraint Satisfaction Problems - CSP. Free Optimization Problem - FOP Nature-Ispired Computig Hadlig Costraits Dr. Şima Uyar September 2006 Itroductio may practical problems are costraied ot all combiatios of variable values represet valid solutios feasible solutios ifeasible

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

LDA-based Non-negative Matrix Factorization for Supervised Face Recognition

LDA-based Non-negative Matrix Factorization for Supervised Face Recognition 1294 JOURNAL OF SOFTWARE, VOL. 9, NO. 5, MAY 2014 LDA-based No-egative Matrix Factorizatio for Supervised Face Recogitio Yu Xue a, Chog Sze Tog b, Jig Yu Yua c a School of Physics ad Telecommuicatio Egieerig,

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

Solving Fuzzy Assignment Problem Using Fourier Elimination Method

Solving Fuzzy Assignment Problem Using Fourier Elimination Method Global Joural of Pure ad Applied Mathematics. ISSN 0973-768 Volume 3, Number 2 (207), pp. 453-462 Research Idia Publicatios http://www.ripublicatio.com Solvig Fuzzy Assigmet Problem Usig Fourier Elimiatio

More information

Feature Selection for Change Detection in Multivariate Time-Series

Feature Selection for Change Detection in Multivariate Time-Series Feature Selectio for Chage Detectio i Multivariate Time-Series Michael Botsch Istitute for Circuit Theory ad Sigal Processig Techical Uiversity Muich 80333 Muich, Germay Email: botsch@tum.de Josef A. Nossek

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

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

Authenticating the Identity of Computer Users with Typing Biometrics and the Fuzzy Min-Max Neural Network

Authenticating the Identity of Computer Users with Typing Biometrics and the Fuzzy Min-Max Neural Network IJBSCHS (2009-4-0-06) Biomedical Soft Computig ad Huma Scieces, Vol.4, No., pp.47-53 (2009) [Origial article] Copyright 995 Biomedical Fuzzy Systems Associatio (Accepted o 2008.0.03) Autheticatig the Idetity

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

x x 2 x Iput layer = quatity of classificatio mode X T = traspositio matrix The core of such coditioal probability estimatig method is calculatig the

x x 2 x Iput layer = quatity of classificatio mode X T = traspositio matrix The core of such coditioal probability estimatig method is calculatig the COMPARATIVE RESEARCHES ON PROBABILISTIC NEURAL NETWORKS AND MULTI-LAYER PERCEPTRON NETWORKS FOR REMOTE SENSING IMAGE SEGMENTATION Liu Gag a, b, * a School of Electroic Iformatio, Wuha Uiversity, 430079,

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

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

On the Accuracy of Vector Metrics for Quality Assessment in Image Filtering

On the Accuracy of Vector Metrics for Quality Assessment in Image Filtering 0th IMEKO TC4 Iteratioal Symposium ad 8th Iteratioal Workshop o ADC Modellig ad Testig Research o Electric ad Electroic Measuremet for the Ecoomic Uptur Beeveto, Italy, September 5-7, 04 O the Accuracy

More information

Eigenimages. Digital Image Processing: Bernd Girod, 2013 Stanford University -- Eigenimages 1

Eigenimages. Digital Image Processing: Bernd Girod, 2013 Stanford University -- Eigenimages 1 Eigeimages Uitary trasforms Karhue-Loève trasform ad eigeimages Sirovich ad Kirby method Eigefaces for geder recogitio Fisher liear discrimat aalysis Fisherimages ad varyig illumiatio Fisherfaces vs. eigefaces

More information

Cubic Polynomial Curves with a Shape Parameter

Cubic Polynomial Curves with a Shape Parameter roceedigs of the th WSEAS Iteratioal Coferece o Robotics Cotrol ad Maufacturig Techology Hagzhou Chia April -8 00 (pp5-70) Cubic olyomial Curves with a Shape arameter MO GUOLIANG ZHAO YANAN Iformatio ad

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

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

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

Journal of Chemical and Pharmaceutical Research, 2013, 5(12): Research Article

Journal of Chemical and Pharmaceutical Research, 2013, 5(12): Research Article Available olie www.jocpr.com Joural of Chemical ad Pharmaceutical Research, 2013, 5(12):745-749 Research Article ISSN : 0975-7384 CODEN(USA) : JCPRC5 K-meas algorithm i the optimal iitial cetroids based

More information

An Efficient Image Rectification Method for Parallel Multi-Camera Arrangement

An Efficient Image Rectification Method for Parallel Multi-Camera Arrangement Y.-S. Kag ad Y.-S. Ho: A Efficiet Image Rectificatio Method for Parallel Multi-Camera Arragemet 141 A Efficiet Image Rectificatio Method for Parallel Multi-Camera Arragemet Yu-Suk Kag ad Yo-Sug Ho, Seior

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

Exploiting Adaptive Packet-Sampling Measurements for Multimedia Traffic Classification

Exploiting Adaptive Packet-Sampling Measurements for Multimedia Traffic Classification Joural of Commuicatios Vol. 9 No. 1 December 014 Exploitig Adaptive Packet-Samplig Measuremets for Multimedia raffic Classificatio Rogqiag Li Ou Li Qig Li ad Ku Dai Natioal Digital Switchig System Egieerig

More information

FEATURES VECTOR FOR PERSONAL IDENTIFICATION BASED ON IRIS TEXTURE

FEATURES VECTOR FOR PERSONAL IDENTIFICATION BASED ON IRIS TEXTURE FEATURES VECTOR FOR PERSONAL IDENTIFICATION BASED ON IRIS TEXTURE R. P. Moreo Departameto de Egeharia Elétrica EESC - USP Av. Trabalhador Sãocarlese, 400 São Carlos / SP Brasil raphael@digmotor.com.br

More information

Extending The Sleuth Kit and its Underlying Model for Pooled Storage File System Forensic Analysis

Extending The Sleuth Kit and its Underlying Model for Pooled Storage File System Forensic Analysis Extedig The Sleuth Kit ad its Uderlyig Model for Pooled File System Foresic Aalysis Frauhofer Istitute for Commuicatio, Iformatio Processig ad Ergoomics Ja-Niclas Hilgert* Marti Lambertz Daiel Plohma ja-iclas.hilgert@fkie.frauhofer.de

More information

BAYESIAN WITH FULL CONDITIONAL POSTERIOR DISTRIBUTION APPROACH FOR SOLUTION OF COMPLEX MODELS. Pudji Ismartini

BAYESIAN WITH FULL CONDITIONAL POSTERIOR DISTRIBUTION APPROACH FOR SOLUTION OF COMPLEX MODELS. Pudji Ismartini Proceedig of Iteratioal Coferece O Research, Implemetatio Ad Educatio Of Mathematics Ad Scieces 014, Yogyakarta State Uiversity, 18-0 May 014 BAYESIAN WIH FULL CONDIIONAL POSERIOR DISRIBUION APPROACH FOR

More information

FUZZY VAULT FUSION BASED MULTIMODAL BIOMETRIC HUMAN RECOGNITION SYSTEM WITH FINGERPRINT AND EAR

FUZZY VAULT FUSION BASED MULTIMODAL BIOMETRIC HUMAN RECOGNITION SYSTEM WITH FINGERPRINT AND EAR FUZZY VAULT FUSION BASED MULTIMODAL BIOMETRIC HUMAN RECOGNITION SYSTEM WITH FINGERPRINT AND EAR R.VINOTHKANNA, 2 Dr.AMITABH WAHI Associate Professor, Departmet of Electroics ad Commuicatio Egieerig, Aapooraa

More information

The Magma Database file formats

The Magma Database file formats The Magma Database file formats Adrew Gaylard, Bret Pikey, ad Mart-Mari Breedt Johaesburg, South Africa 15th May 2006 1 Summary Magma is a ope-source object database created by Chris Muller, of Kasas City,

More information

Improving Information Retrieval System Security via an Optimal Maximal Coding Scheme

Improving Information Retrieval System Security via an Optimal Maximal Coding Scheme Improvig Iformatio Retrieval System Security via a Optimal Maximal Codig Scheme Dogyag Log Departmet of Computer Sciece, City Uiversity of Hog Kog, 8 Tat Chee Aveue Kowloo, Hog Kog SAR, PRC dylog@cs.cityu.edu.hk

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

Data Warehousing. Paper

Data Warehousing. Paper Data Warehousig Paper 28-25 Implemetig a fiacial balace scorecard o top of SAP R/3, usig CFO Visio as iterface. Ida Carapelle & Sophie De Baets, SOLID Parters, Brussels, Belgium (EUROPE) ABSTRACT Fiacial

More information

Comparison of classification algorithms in the task of object recognition on radar images of the MSTAR base

Comparison of classification algorithms in the task of object recognition on radar images of the MSTAR base Compariso of classificatio algorithms i the task of object recogitio o radar images of the MSTAR base A.A. Borodiov 1, V.V. Myasikov 1,2 1 Samara Natioal Research Uiversity, 34 Moskovskoe Shosse, 443086,

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

. 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