AN APPLICATION OF THE TCRBF NEURAL NETWORK IN MULTI-NODE FAULT DIAGNOSIS METHOD
|
|
- Karin Nash
- 5 years ago
- Views:
Transcription
1 XIX IMEKO World Congress Fundamental and Appled Metrology September 6 11, 2009, Lsbon, Portugal AN APPLICATION OF THE TCRBF NEURAL NETWORK IN MULTI-NODE FAULT DIAGNOSIS METHOD Zbgnew Czaja 1, Mchał Kowalewsk 2 1 Gdansk Unversty of Technology, Faculty of Electroncs, Telecommuncatons and Informatcs, Department of Optoelectroncs and Electronc Systems, Gdansk, Poland, zbczaja@pg.gda.pl 2 Gdansk Unversty of Technology, Faculty of Electroncs, Telecommuncatons and Informatcs, Department of Optoelectroncs and Electronc Systems, Gdansk, Poland, Mchal.Kowalewsk@et.pg.gda.pl Abstract Ths paper presents the new self-testng method for dagnoss of analog parts n mxed-sgnal embedded systems controlled by mcrocontrollers. The tested analog part s stmulated by a snus-wave suppled by the onboard generator and ts responses are sampled n selected nodes by mcrocontrollers ADC. The measurement space s represented by dfferences between values of selected node voltages. Fault detecton and localzaton s performed by a Two-Center Radal Bass Functon (TCRBF) Neural Network. The dagnoss procedure was mplemented and smulated n a PC. Keywords fault dagnoss, neural network, BIST 1. INTRODUCTION At present, embedded electronc systems whch are characterzed by an ntellgent unt, often based on a mcrocontroller, a dgtal sgnal processor (DSP) or a programmable devce (e.g. FPGA, CPLD), predomnate on the electronc market, because nformaton about the operatonal envronment and controlled objects are often obtaned va analog sensors. Analog sgnals are transmtted and ntally processed n analog parts, however analog-todgtal processng and data processng are realzed by dgtal parts. Therefore, the analog part has to work correctly, because an ncorrect measurement sgnal can lead to a wrong decson of the control unt, what can even result n damage of the controlled devce. Hence, the embedded system should be able to run self-testng of analog parts. Self-testng of analog crcuts bases on fault dagnoss methods. When elaboratng these methods we have to take nto consderaton contnuous values of crcut elements, the contnuous nature of stmulaton and response sgnals and the fact that these sgnals can assume the shape of any functon, the presence of element tolerances and crcut nonlneartes. Addtonally, especally for non-electrcal objects modelled by electrcal crcuts we have only poorly defned system models. Thus neural networks can be a very effcent soluton n fault dagnoss methods. Many types of neural networks are used as fault classfers: backpropagaton [1], probablstc [2], self-organzng [3] radal bass functon [4,5], neural networks based on TCRB Functons [6]. In ths paper a new self-testng procedure wll be presented. It bases on a new mult-port fault dagnoss method and the smplfed TCRBF neural classfer [7]. The dagnoss method s used to create a fault dctonary n the form of dspersed localzaton curves deally suted for the TCRBF neural classfer. Addtonally, measurements of parameters of the tested analog part are performed by the reconfgurable BIST created from perpheral devces of the control unt controllng the electronc embedded system. 2. FAULT DIAGNOSIS METHOD The method wll be llustrated on an exemplary analog part represented by the low-pass Tow-Tomas flter shown n Fg. 1. Fg. 1. Tested analog crcut low-pass Tow-Tomas flter, where R1 = R2 = R3 = R4 = R5 = R6 = 10 kω, C1 = 33 nf, C2 = 6.8 nf Generally, the self-testng procedure conssts of three stages. At frst, durng the pre-testng stage, the TCRBF neural classfer s constructed on the bass of a famly of dspersed localzaton curves. These curves descrbe the behavour of the tested crcut followng changes of ts elements values. The second stage s responsble for snuswave stmulaton of the tested crcut and for measurements of real and magnary values of voltages n accessble nodes. At the last stage, the TCRBF neural classfer detects and localzes a fault of the analog part Idea of the method Let X N be a set of N accessble measurement nodes X N = {X 1, X 2,..., X N }. Thus, we have the followng set of values of node voltages: u N = {v n,w n } n = 1,..., N, where
2 u n = v n + j w n. These voltages can drectly create a (2 N)- dmensonal measurement space n a way smlar to [9,10], but n ths case localzaton curves representng devatons of values of crcut elements often characterze wdely dfferent lengths, whch depends on crcut senstvty, and often they are not stuated optmally. Ths makes t dffcult to perform correct fault localzaton. Hence, t s proposed to create a new measurement space, where coordnates have the followng forms: {v n,n-1, w n,n-1 } n = 2,.., N, where v n,n-1, = v n - v n-1, w n,n-1, = w n - w n-1. That s we create the space represented by dfferences between values of node voltages. In ths case we mprove the dynamcs of length changes of localzaton curves. Lengths are more smlar between themselves and curves are more separated as shown n Fg. 2 and Fg. 3. It wll be explaned on the example of the crcut shown n Fg. 1. The denomnator for all ts node transfer functons s followng: ( R R R R X R R R + X X R ) d = R (1) R6 where X 1 = -1/(jω C 1 ), X 2 = -1/(jω C 2 ). It s dependent on all elements, especally on R 1. However, the numerators for next node transfer functons have the forms: Thus, we obtan the followng dfferences between numerators of transfer functons for respectve nodes: (2) n two three-dmensonal spaces (Fg. 2, 3). Localzaton curves n these fgures were drawn for ±50% changes of element values wth reference to ther nomnal values p j nom. Fg. 2. The map of localzaton curves for the tested crcut n the 3-dmensonal measurement space: Re(u 2,1 ), Im(u 2,1 ), Re(u 3,2 ). Localzaton curves are a graphcal descrpton of crcut propretes followng from changes of ts element values. All curves cross at a pont whch s the nomnal state of the crcut. Assumpton of element tolerances causes dsperson of localzaton curves. Thus we need a classfer wth good generalzaton capabltes to correctly localze faults. Ths requrement can be fulflled by the TCRBF classfer, whch s constructed on the bass of dspersed localzaton curves. It s seen from (2) and (3) that the dfference m 2 m 1 between the 1 st and the 2 nd node functon numerators s dependent on the same elements as the respectve numerators. The same stuaton s for the m 3 m 2 dfference. That s we decrease the sze of the measurement space what decreases the sze of the fault dctonary, smultaneously keepng the same level of the localzaton resoluton. But ths operaton ncreases the dynamc of measurement sgnals, because sgnals for neghbourng nodes are moved about 180 and they added to themselves (see Fg. 1). The new measurement space s descrbed by the followng transformaton: (3) T ( p ) N = ( n= 2 Re + Im ( u ( p ) u ( p )) n n 1 ( u ( p ) u ( p )) ) n n 1 2n 2 2n 3 where: n - s a coordnate vector along the n axs, u n (p ) complex value of voltage n the node n for a change of p value of the -th element, = 1,..., I, I the number of crcut elements, N the number of accessble measurement nodes. The transformaton (4) maps the changes of crcut element values {p } =1,,I nto a famly of localzaton curves placed n the (2 N 2)-dmensonal measurement space. For the crcut shown n Fg. 1 we have a 4- dmensonal space. Thus the transformaton (1) s presented (4) Fg. 3. The map of localzaton curves for the tested crcut n the 3-dmensonal measurement space: Re(u 2,1 ), Im(u 2,1 ), Im(u 3,2 ) The measurement procedure The measurement procedure s performed by mcrocontroller ATmega16 and ts nternal devces [11]: analog multplexer, 10-bt ADC and 16-bt Tmer 1 (Fg. 4). The tested analog part s stmulated by the snus-wave and ts responses are sampled n partcular nodes by the ADC n moments establshed by the Tmer.
3 vn = wn = ( un,1 un,1 + un,2 un,2 ) ( u u u u ) n,1 n,2 n,2 n,1 where the value χ = 1 / U n s constant and known, what consderably smplfes calculatons of these values, because we use only addton and multplcaton operatons. Next t calculates dfferences between these values: v n,n-1, = v n - v n-1, w n,n-1, = w n - w n-1. χ χ (5) Fg. 4. Example of the electronc embedded system n the selftestng confguraton Tmngs of the measurement procedure are shown n Fg. 5. Each sgnal s sampled three tmes, where tme dstances between samples are set to one fourth of the perod T. Tme dstances between samples for subsequent sgnals are equal to a half of the perod. Thanks to ths, t s allowed to sample all sgnals at the same moments n relaton to the start of samplng, because samplng of the next sgnal s shfted about one perod. Samplng of the nput sgnal u n s needed to establsh a random shft tme T α of the samplng seres. The thrd sample of each sgnal s used to elmnate the voltage offset. The frst u n,1 and second u n,2 samples of the node voltage sgnal u n are used to calculate ther real and magnary values. Fg. 6. The algorthm of the measurement procedure 3. USE OF THE TCRBF NEURAL NETWORK Fg. 5. Tmngs of the measured sgnals durng the self-testng procedure n the nput node and nodes X 1 and X 2. The algorthm of the measurement procedure (Fg. 6) conssts of a measurement part whch s responsble for control of the measurement. The man functon confgures the ADC, starts the tmer and wats for the end of samplng (t wats for the measurement of nne voltage samples). The nterrupt servce of the tmer starts the ADC converson and actualzes the countng tme between next samples. The nterrupt servce of the ADC converson complete saves the voltage samples and changes the channel of the analog multplexer to measure the next node voltage. Durng the calculaton part, the mcrocontroller computes the real v n and magnary w n parts of the n-th node voltage basng on: Detecton and localzaton of faults n the proposed method are performed by a neural network wth TCRB functons [6,7]. The usefulness of the TCRB functon n dctonary methods based on localzaton curves follows from the fact that the localzaton curve (Fg. 7a) can be transformed only wth a few TCRB functons (Fg. 7d). In advance, some experments wth Radal Bass Functon Neural Networks (RBFNN) for dagnoss purposes were performed [4,5]. Unfortunately an accurate transformaton of localzaton curves wth RBFNN requres a lot of Radal Bass (RB) functons n the archtecture (Fg. 7b). Hence t appears an dea to construct a new neuron for better transformaton of stretched clusters of dspersed localzaton curves. The TCRB functon radally maps the space around a lne segment wth endponts c (1) and c (2) (Fg. 7c) wth the followng equaton: n 1 2 y( x ) = exp ( ( )) ( ) x w x, (6) 2 s x = 1
4 where: s(x) s the scalng functon descrbng the localzaton curve dsperson changng form σ 1 to σ 2 and w (x) ( = 1, 2,, n) are center functons dependng on coordnates of centers c (1) and c (2) [6]. real values M y 1 v 2,1 w 2,1 M 2 y 2 v 3,2 f 1 f 2 f 0,1 - values w 3,2 Input layer 1 st hdden layer wth TCRB functons M C y C λ 2 nd hdden layer wth maxmum functons Output layer Indcaton faulty component f C Indcaton state of the crcut s Fg. 9. The TCRBF neural network structure. Fg. 7. Idea of localzaton curve transformaton wth RB and TCRB functons. Applyng some TCRB functons to each localzaton curve of the testng crcut gves the possblty to obtan nformaton about dstances between the measurement pont and localzaton curves. Ths nformaton s gven as a vector of real values from the range (0, 1]. If the measurement pont les near the localzaton curve then one of TCRB functons assgned to that curve has the value close to 1. Otherwse, f the measurement pont les far from the localzaton curve, all TCRB functons assgned to ths curve have values close to 0. The number of requred TCRB functons for each localzaton curve depends on ts curvature. For a curve smlar to a straght lne only one TCRBF s needed. The more bent the curve s the more TCRB functons are requred. A graphcal llustraton of actvaton regons of TCRB functons for exemplary dspersed localzaton curves s shown n Fg. 8. Fg. 8. Graphcal llustraton of actvaton regons of TCRB functons for exemplary dspersed localzaton curves Archtecture of TCRBF classfer The TCRBF classfer proposed n ths paper s a threelayer feed-forward network (Fg. 9). Its archtecture s based on other neural networks wth RB functons (RBFNN, Probablstc, Generalsed Regresson). TCRB functons assgned to localzaton curves are placed n the 1 st hdden layer. They play the role of radal mappng of dspersed localzaton curves to real values from the range (0, 1]. Dsperson s caused by crcut element tolerances. Neurons n the 2 nd hdden layer group TCRB functons n C classes, whch nclude sngletons and ambguty groups, and produce maxmum values (M) of TCRBF outputs. The output layer s a perceptron wth a hard lmt transfer functon, whch produces a vector f wth 0 and 1 values. Each neuron n the output layer s fed only wth two sgnals: maxmum value of TCRB functons assgned to a class and constant bas λ. The network output vector f combnes nformaton about the state of the testng crcut and moreover about the fault class. An addtonal summng functon sums perceptron output values and gves the value s = f 1 + f f C, whch presents the state of the testng crcut as a sngle nteger value from the range [0, C]. The TCRBF classfer s able to dstngush between three states of the crcut: fault-free state (s = C), sngle fault or ambguty group (1 s < C), multple fault (s = 0). A. Fault free state If the measurement pont s located near the nomnal pont then all TCRB functons located closest to the nomnal pont and assgned to dfferent classes have output values y 1,, y C greater than λ. Hence the sum of all values from the perceptron outputs gves s = C. It ndcates a fault-free state of the crcut. B. Sngle fault or ambguty group Else f the measurement pont s located near one of localzaton curves assgned to a fault class and far from others, then only one value y s grater than λ and the sum of all values from the perceptron outputs gves s = 1. In ths case a sngle one n the output vector f ndcates a fault class (sngle fault or ambguty group). We meet wth ambguty also when 1 < s < C. In ths case the measurement pont s located near more than one localzaton curves assgned to dfferent fault classes.
5 C. Multple fault Fnally f the measurement pont s located far from any of localzaton curves then all values y are lower than λ and the sum of all values from the perceptron outputs gves s = 0. In ths case we meet wth a multple fault and t s mpossble to ndcate whch element s faulty. A constant parameter λ (0, 1), appled to the perceptron layer, makes t possble to descrbe the szes of decson regons n the measurement space. For λ close to 1 decson regons are narrowed to spaces located near localzaton curves. It s best suted for crcuts wth small element tolerances. Otherwse, f λ s close to 0, decson regons are wde and crcuts wth greater element tolerances can be correctly dagnosed Constructon of TCRBF classfer In contrast to RBF or feed-forward back-propagaton (FFBP) neural networks, TCRBFNN does not requre tranng. Centers of TCRB functons are placed on localzaton curves of the testng crcut wth a known model. Scalng parameters of TCRB functons are obtaned on the bass of Monte Carlo analyss n selected centers. The center selecton method bases on lnear nterpolaton of localzaton curves. Startng from two centers placed on extreme ponts of the localzaton curve, n next steps ponts farthest to the nterpolaton curve are selected as new centers and also as new nterpolaton nodes. Ths procedure s repeated untl the maxmum dstances between localzaton and nterpolaton curves are lower then the assumed dstance d max. In the second step for all selected centers the Monte Carlo analyss s performed n order to obtan localzaton curves dsperson. Wth assumpton of normal dstrbuton of ponts n each cluster, standard devatons of multvarate normal dstrbuton are estmated and used as scalng parameters of TCRB functons. Fg. 10 presents nterpolaton curves and data clusters obtaned from the Monte Carlo analyss wth assumpton of 3% element tolerances for the crcut shown n Fg. 1. bnd dentfcaton curves wth fault classes. In case of lack of the crcut model, when only expermentally acqured measurements are avalable, TCRBF centers can be obtaned wth the fuzzy c-means clusterng algorthm [4]. The parameter λ of the output layer s selected expermentally n order to decrease the classfcaton error Smulaton results The proposed dagnoss procedure wth the TCRBF classfer was used for faults detecton and localzaton n the low-pass flter shown n Fg. 1. Three TCRBF classfers wth dfferent archtectures were constructed (Table 1) and ther generalzaton capabltes were compared. A dfferent number of the followng nput coordnates was used n each classfer: Re(u 2,1 ), Im(u 2,1 ), Re(u 3,2 ), Im(u 3,2 ). Input dmenson Table 1. TCRBF classfers parameters. Number of TCRB functons Network archtecture Number of TCRB functons per each element R 1 R 2 R 3 R 4 R 5 R 6 C 1 C 2 2D D D The number of TCRB functons and classes n each classfer depends on the dmenson of the nput space (except the curvature of localzaton curves). For more dmensons, localzaton curves are more dstant from each other and faults are more dstngushable. In ths case more TCRB functons are needed and more classes can be created. Localzaton curves of some elements cover each other and form ambguty groups, for example {R 4, C 2 }. Hence one needs to apply TCRB functons only for one element n the ambguty group. An ncrease n the number of nput dmensons from 2 to 3 caused separaton of localzaton curves n the ambguty group {R 3, R 5, R 6, C 1 }. For ths group two classes were formed: {R 3,C 1 } and {R 5, R 6 }. Graphcal llustratons of the 1 st hdden layer (wth TCRB functons) actvaton regons n 2- and 3-dmensonal nput spaces are shown n Fg. 11 and Fg. 12. We can see a very good ft to localzaton curves. Fg. 10. Interpolaton curves and data clusters obtaned for the crcut shown n Fg. 1 and used n TCRBF classfer constructon stage. Connectons between the 1 st and the 2 nd hdden layer are made easly, as we know the crcut model and know how to Fg. 11. Graphcal transformaton of dspersed localzaton curves by the TCRBF neural classfer n the 2-dmensonal measurement space wth coordnates: Re(u 2,1 ), Im(u 2,1 ).
6 4. CONCLUSIONS Fg. 12. Graphcal transformaton of dspersed localzaton curves by the TCRBF neural classfer n the 3-dmensonal measurement space wth coordnates: Re(u 2,1 ), Im(u 2,1 ), Re(u 3,2 ). The classfcaton error of the created TCRBF classfers dd not exceed 2% and was manly caused by mproperly classfed sgnatures near the nomnal state of the crcut. The TCRBF classfer constructed on a data set belongng to dspersed localzaton curves s able to dstngush between the nomnal state of the crcut, sngle faults and multple faults (Fg. 13). Ths feature can be acheved wth the RBF neural network, but as mentoned earler, a correct transformaton of localzaton curves requres a sgnfcant ncrease n the number of RB functons. Fg. 13. Decson regons of the TCRBF classfer wth the archtecture n the 2-dmensonal measurement space for λ = 0.5. In the case of the FFBP neural network wth the archtecture , traned on the same data set as the TCRBF classfer, the classfcaton error on the test set exceeds 30%. It follows from the fact that neurons n FFBP network dvde the pattern space usng hyperplanes. Hence decson regons of that network go beyond dspersed localzaton curves and extend to nfnty. To decrease ths effect the FFBP network requres addton of a multple faults data set n the tranng data set. Addtonally the number of samples of the tranng set assgned to dspersed localzaton curves must be ncreased. Ths consderably extends the classfer constructon procedure wth reference to the TCRBF neural network. The new self-testng method wth the TCRBF classfer presented n ths paper s an effcent way to dagnose analog parts n mxed-sgnal embedded systems controlled by mcrocontrollers. The novelty of the method les n creatng a new measurement space represented by dfferences between values of node voltages. The TCRBF classfer used to detect the state of the crcut, as opposed to other types of neural classfers (back-propagaton, radal bass, probablstc), has generalzaton capabltes deally suted for dctonary methods based on localzaton curves. Unlke FFBP and RBF networks, the TCRBFNN requres a lower sze of the data set used for constructon. Hence the constructon phase s shorter and less burdensome. Furthermore, on the bass of the lmted sze of the constructon data set, the TCRBF classfer gves the possblty to extend dagnostc nformaton about the testng crcut. It s possble to dstngush between a fault-free state of the crcut, a sngle fault, an ambguty group and a multple fault. REFERENCES [1] Czaja Z., Zelonko R. Fault dagnoss n electronc crcuts based on blnear transformaton n 3-D and 4-D spaces, Trans. on Instr. and Measurement, vol. 52, nº 1, pp , [2] Yang Z. R., Zwolnsk M. Applyng a robust heteroscedastc probablstc neural network to analog fault detecton and classfcaton, IEEE Trans. CAD IC&S, vol. 19, pp , [3] Collns P., Yu S., Eckersall K. R., Jervs B. W., Bell I. M., Taylor G. E. Applcaton of Kohonen and supervsed forced organzaton maps to fault dagnoss of electronc analog crcuts, Electronc Letters, vol. 30, nº 22, pp , [4] Catelan M., Fort A. Fault dagnoss of electronc analog crcuts usng a radal bass functon network classfer, Measurement, vol. 28, Issue 3, pp , [5] Toczek W., Kowalewsk M. A neural network based system for soft faults dagnoss n electronc crcuts, Metrology and Measurements Systems, vol. 12, nº 4, pp , [6] Kowalewsk M. Two-center radal bass functon network for classfcaton of soft faults n electronc analog crcuts, In proc. of the IMTC 2007 Conference, Warsaw, Poland, [7] Czaja Z., Kowalewsk M. A new method for dagnoss of analog parts n electronc embedded systems wth two center radal bass functon neural networks, 16 th IMEKO TC4 Symposum, pp , Italy, Florence, September, [8] Czaja Z., A dagnoss method of analog parts of mxed-sgnal systems controlled by mcrocontrollers, Measurement, vol. 40, nº 2, pp , [9] Czaja Z., Usng a square-wave sgnal for fault dagnoss of analog parts of mxed-sgnal electronc embedded systems, IEEE Transactons on Instrumentaton and Measurement, vol. 57, nº. 8, pp , August, [10] Czaja Z., A fault dagnoss algorthm of analog crcuts based on node-voltage relaton, 12th IMEKO TC1-TC7 Jont Symposum, pp , Annecy, France, September, [11] Atmel Corporaton, 8-bt AVR mcrocontroller wth 16k Bytes In-System Programmable Flash, ATmega16, ATmega16L, PDF fle, Avalable from:
Machine Learning 9. week
Machne Learnng 9. week Mappng Concept Radal Bass Functons (RBF) RBF Networks 1 Mappng It s probably the best scenaro for the classfcaton of two dataset s to separate them lnearly. As you see n the below
More informationClassifier Selection Based on Data Complexity Measures *
Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.
More informationClassifying Acoustic Transient Signals Using Artificial Intelligence
Classfyng Acoustc Transent Sgnals Usng Artfcal Intellgence Steve Sutton, Unversty of North Carolna At Wlmngton (suttons@charter.net) Greg Huff, Unversty of North Carolna At Wlmngton (jgh7476@uncwl.edu)
More informationParallelism for Nested Loops with Non-uniform and Flow Dependences
Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr
More informationy and the total sum of
Lnear regresson Testng for non-lnearty In analytcal chemstry, lnear regresson s commonly used n the constructon of calbraton functons requred for analytcal technques such as gas chromatography, atomc absorpton
More informationWishing you all a Total Quality New Year!
Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma
More informationImprovement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration
Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,
More informationProgramming in Fortran 90 : 2017/2018
Programmng n Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Exercse 1 : Evaluaton of functon dependng on nput Wrte a program who evaluate the functon f (x,y) for any two user specfed values
More informationA Fast Content-Based Multimedia Retrieval Technique Using Compressed Data
A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,
More informationLecture 5: Multilayer Perceptrons
Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented
More informationA Binarization Algorithm specialized on Document Images and Photos
A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a
More information3D vector computer graphics
3D vector computer graphcs Paolo Varagnolo: freelance engneer Padova Aprl 2016 Prvate Practce ----------------------------------- 1. Introducton Vector 3D model representaton n computer graphcs requres
More informationSLAM Summer School 2006 Practical 2: SLAM using Monocular Vision
SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,
More informationEdge Detection in Noisy Images Using the Support Vector Machines
Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona
More informationSteps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices
Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between
More informationA Fast Visual Tracking Algorithm Based on Circle Pixels Matching
A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng
More informationSome Advanced SPC Tools 1. Cumulative Sum Control (Cusum) Chart For the data shown in Table 9-1, the x chart can be generated.
Some Advanced SP Tools 1. umulatve Sum ontrol (usum) hart For the data shown n Table 9-1, the x chart can be generated. However, the shft taken place at sample #21 s not apparent. 92 For ths set samples,
More informationDetection of an Object by using Principal Component Analysis
Detecton of an Object by usng Prncpal Component Analyss 1. G. Nagaven, 2. Dr. T. Sreenvasulu Reddy 1. M.Tech, Department of EEE, SVUCE, Trupath, Inda. 2. Assoc. Professor, Department of ECE, SVUCE, Trupath,
More informationVRT012 User s guide V0.1. Address: Žirmūnų g. 27, Vilnius LT-09105, Phone: (370-5) , Fax: (370-5) ,
VRT012 User s gude V0.1 Thank you for purchasng our product. We hope ths user-frendly devce wll be helpful n realsng your deas and brngng comfort to your lfe. Please take few mnutes to read ths manual
More informationOutline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1
4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:
More informationR s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes
SPH3UW Unt 7.3 Sphercal Concave Mrrors Page 1 of 1 Notes Physcs Tool box Concave Mrror If the reflectng surface takes place on the nner surface of the sphercal shape so that the centre of the mrror bulges
More informationSupport Vector Machines
/9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.
More informationAn Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation
17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 An Iteratve Soluton Approach to Process Plant Layout usng Mxed
More informationThe Codesign Challenge
ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.
More informationSubspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;
Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features
More information2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements
Module 3: Element Propertes Lecture : Lagrange and Serendpty Elements 5 In last lecture note, the nterpolaton functons are derved on the bass of assumed polynomal from Pascal s trangle for the fled varable.
More informationSix-Band HDTV Camera System for Color Reproduction Based on Spectral Information
IS&T's 23 PICS Conference Sx-Band HDTV Camera System for Color Reproducton Based on Spectral Informaton Kenro Ohsawa )4), Hroyuk Fukuda ), Takeyuk Ajto 2),Yasuhro Komya 2), Hdeak Hanesh 3), Masahro Yamaguch
More informationComparison Study of Textural Descriptors for Training Neural Network Classifiers
Comparson Study of Textural Descrptors for Tranng Neural Network Classfers G.D. MAGOULAS (1) S.A. KARKANIS (1) D.A. KARRAS () and M.N. VRAHATIS (3) (1) Department of Informatcs Unversty of Athens GR-157.84
More informationEvaluation of an Enhanced Scheme for High-level Nested Network Mobility
IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.15 No.10, October 2015 1 Evaluaton of an Enhanced Scheme for Hgh-level Nested Network Moblty Mohammed Babker Al Mohammed, Asha Hassan.
More informationDistance Calculation from Single Optical Image
17 Internatonal Conference on Mathematcs, Modellng and Smulaton Technologes and Applcatons (MMSTA 17) ISBN: 978-1-6595-53-8 Dstance Calculaton from Sngle Optcal Image Xao-yng DUAN 1,, Yang-je WEI 1,,*
More informationBOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET
1 BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET TZU-CHENG CHUANG School of Electrcal and Computer Engneerng, Purdue Unversty, West Lafayette, Indana 47907 SAUL B. GELFAND School
More informationTECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z.
TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS Muradalyev AZ Azerbajan Scentfc-Research and Desgn-Prospectng Insttute of Energetc AZ1012, Ave HZardab-94 E-mal:aydn_murad@yahoocom Importance of
More informationAn Image Fusion Approach Based on Segmentation Region
Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua
More informationAdvanced Computer Networks
Char of Network Archtectures and Servces Department of Informatcs Techncal Unversty of Munch Note: Durng the attendance check a stcker contanng a unque QR code wll be put on ths exam. Ths QR code contans
More informationOutline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:
Self-Organzng Maps (SOM) Turgay İBRİKÇİ, PhD. Outlne Introducton Structures of SOM SOM Archtecture Neghborhoods SOM Algorthm Examples Summary 1 2 Unsupervsed Hebban Learnng US Hebban Learnng, Cntd 3 A
More informationTerm Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task
Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto
More informationHelsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr)
Helsnk Unversty Of Technology, Systems Analyss Laboratory Mat-2.08 Independent research projects n appled mathematcs (3 cr) "! #$&% Antt Laukkanen 506 R ajlaukka@cc.hut.f 2 Introducton...3 2 Multattrbute
More informationVirtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory
Background EECS. Operatng System Fundamentals No. Vrtual Memory Prof. Hu Jang Department of Electrcal Engneerng and Computer Scence, York Unversty Memory-management methods normally requres the entre process
More informationVISUAL SELECTION OF SURFACE FEATURES DURING THEIR GEOMETRIC SIMULATION WITH THE HELP OF COMPUTER TECHNOLOGIES
UbCC 2011, Volume 6, 5002981-x manuscrpts OPEN ACCES UbCC Journal ISSN 1992-8424 www.ubcc.org VISUAL SELECTION OF SURFACE FEATURES DURING THEIR GEOMETRIC SIMULATION WITH THE HELP OF COMPUTER TECHNOLOGIES
More informationParallel Inverse Halftoning by Look-Up Table (LUT) Partitioning
Parallel Inverse Halftonng by Look-Up Table (LUT) Parttonng Umar F. Sddq and Sadq M. Sat umar@ccse.kfupm.edu.sa, sadq@kfupm.edu.sa KFUPM Box: Department of Computer Engneerng, Kng Fahd Unversty of Petroleum
More informationSimulation Based Analysis of FAST TCP using OMNET++
Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months
More informationHierarchical clustering for gene expression data analysis
Herarchcal clusterng for gene expresson data analyss Gorgo Valentn e-mal: valentn@ds.unm.t Clusterng of Mcroarray Data. Clusterng of gene expresson profles (rows) => dscovery of co-regulated and functonally
More informationSimple March Tests for PSF Detection in RAM
Smple March Tests for PSF Detecton n RAM Ireneusz Mroze Balysto Techncal Unversty Computer Scence Department Wejsa 45A, 5-35 Balysto POLAND mroze@.pb.balysto.pl Eugena Buslowsa Balysto Techncal Unversty
More informationAUTOMATED METHOD FOR STATISTICAL PROCESSING OF AE TESTING DATA
AUTOMATED METHOD FOR STATISTICAL PROCESSING OF AE TESTING DATA V. A. Barat and A. L. Alyakrtsky Research Dept, Interuns Ltd., bld. 24, corp 3-4, Myasntskaya str., Moscow, 0000, Russa Keywords: sgnal processng,
More informationTN348: Openlab Module - Colocalization
TN348: Openlab Module - Colocalzaton Topc The Colocalzaton module provdes the faclty to vsualze and quantfy colocalzaton between pars of mages. The Colocalzaton wndow contans a prevew of the two mages
More informationAn Improved Neural Network Algorithm for Classifying the Transmission Line Faults
1 An Improved Neural Network Algorthm for Classfyng the Transmsson Lne Faults S. Vaslc, Student Member, IEEE, M. Kezunovc, Fellow, IEEE Abstract--Ths study ntroduces a new concept of artfcal ntellgence
More informationProblem Definitions and Evaluation Criteria for Computational Expensive Optimization
Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty
More informationSHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE
SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE Dorna Purcaru Faculty of Automaton, Computers and Electroncs Unersty of Craoa 13 Al. I. Cuza Street, Craoa RO-1100 ROMANIA E-mal: dpurcaru@electroncs.uc.ro
More informationFeature Reduction and Selection
Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components
More informationDiscrimination of Faulted Transmission Lines Using Multi Class Support Vector Machines
16th NAIONAL POWER SYSEMS CONFERENCE, 15th-17th DECEMBER, 2010 497 Dscrmnaton of Faulted ransmsson Lnes Usng Mult Class Support Vector Machnes D.hukaram, Senor Member IEEE, and Rmjhm Agrawal Abstract hs
More informationS1 Note. Basis functions.
S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type
More informationMachine Learning: Algorithms and Applications
14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of
More informationPositive Semi-definite Programming Localization in Wireless Sensor Networks
Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer
More informationFEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur
FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents
More informationThe Shortest Path of Touring Lines given in the Plane
Send Orders for Reprnts to reprnts@benthamscence.ae 262 The Open Cybernetcs & Systemcs Journal, 2015, 9, 262-267 The Shortest Path of Tourng Lnes gven n the Plane Open Access Ljuan Wang 1,2, Dandan He
More information6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour
6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the
More informationFuzzy Logic Based RS Image Classification Using Maximum Likelihood and Mahalanobis Distance Classifiers
Research Artcle Internatonal Journal of Current Engneerng and Technology ISSN 77-46 3 INPRESSCO. All Rghts Reserved. Avalable at http://npressco.com/category/jcet Fuzzy Logc Based RS Image Usng Maxmum
More informationContent Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers
IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth
More informationCompiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz
Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster
More informationA mathematical programming approach to the analysis, design and scheduling of offshore oilfields
17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and
More informationLecture 4: Principal components
/3/6 Lecture 4: Prncpal components 3..6 Multvarate lnear regresson MLR s optmal for the estmaton data...but poor for handlng collnear data Covarance matrx s not nvertble (large condton number) Robustness
More informationX- Chart Using ANOM Approach
ISSN 1684-8403 Journal of Statstcs Volume 17, 010, pp. 3-3 Abstract X- Chart Usng ANOM Approach Gullapall Chakravarth 1 and Chaluvad Venkateswara Rao Control lmts for ndvdual measurements (X) chart are
More informationLoad-Balanced Anycast Routing
Load-Balanced Anycast Routng Chng-Yu Ln, Jung-Hua Lo, and Sy-Yen Kuo Department of Electrcal Engneerng atonal Tawan Unversty, Tape, Tawan sykuo@cc.ee.ntu.edu.tw Abstract For fault-tolerance and load-balance
More informationUsing Fuzzy Logic to Enhance the Large Size Remote Sensing Images
Internatonal Journal of Informaton and Electroncs Engneerng Vol. 5 No. 6 November 015 Usng Fuzzy Logc to Enhance the Large Sze Remote Sensng Images Trung Nguyen Tu Huy Ngo Hoang and Thoa Vu Van Abstract
More informationRADIX-10 PARALLEL DECIMAL MULTIPLIER
RADIX-10 PARALLEL DECIMAL MULTIPLIER 1 MRUNALINI E. INGLE & 2 TEJASWINI PANSE 1&2 Electroncs Engneerng, Yeshwantrao Chavan College of Engneerng, Nagpur, Inda E-mal : mrunalngle@gmal.com, tejaswn.deshmukh@gmal.com
More informationK-means and Hierarchical Clustering
Note to other teachers and users of these sldes. Andrew would be delghted f you found ths source materal useful n gvng your own lectures. Feel free to use these sldes verbatm, or to modfy them to ft your
More informationA PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION
1 THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Seres A, OF THE ROMANIAN ACADEMY Volume 4, Number 2/2003, pp.000-000 A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION Tudor BARBU Insttute
More informationDeep learning is a good steganalysis tool when embedding key is reused for different images, even if there is a cover source-mismatch
Deep learnng s a good steganalyss tool when embeddng key s reused for dfferent mages, even f there s a cover source-msmatch Lonel PIBRE 2,3, Jérôme PASQUET 2,3, Dno IENCO 2,3, Marc CHAUMONT 1,2,3 (1) Unversty
More informationReview of approximation techniques
CHAPTER 2 Revew of appromaton technques 2. Introducton Optmzaton problems n engneerng desgn are characterzed by the followng assocated features: the objectve functon and constrants are mplct functons evaluated
More informationImage Representation & Visualization Basic Imaging Algorithms Shape Representation and Analysis. outline
mage Vsualzaton mage Vsualzaton mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and Analyss outlne mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and
More informationComparison of SVM and ANN for classification of eye events in EEG
J. Bomedcal Scence and Engneerng, 2011, 4, 62-69 do:10.4236/jbse.2011.41008 Publshed Onlne January 2011 (http://www.scrp.org/journal/jbse/). Comparson of SVM and ANN for classfcaton of eye events n EEG
More informationTHE FAULT LOCATION ALGORITHM BASED ON TWO CIRCUIT FUNCTIONS
U THE FAULT LOCATION ALGORITHM BASED ON TWO CIRCUIT FUNCTIONS Z. Czaa Char of Electronc Measurement, Faculty of Electroncs, Telecommuncatons an Informatcs, Techncal Unversty of Gañsk, Polan The paper presents
More informationCluster Analysis of Electrical Behavior
Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School
More informationThe Comparison of Calibration Method of Binocular Stereo Vision System Ke Zhang a *, Zhao Gao b
3rd Internatonal Conference on Materal, Mechancal and Manufacturng Engneerng (IC3ME 2015) The Comparson of Calbraton Method of Bnocular Stereo Vson System Ke Zhang a *, Zhao Gao b College of Engneerng,
More informationClassification / Regression Support Vector Machines
Classfcaton / Regresson Support Vector Machnes Jeff Howbert Introducton to Machne Learnng Wnter 04 Topcs SVM classfers for lnearly separable classes SVM classfers for non-lnearly separable classes SVM
More informationMOTION PANORAMA CONSTRUCTION FROM STREAMING VIDEO FOR POWER- CONSTRAINED MOBILE MULTIMEDIA ENVIRONMENTS XUNYU PAN
MOTION PANORAMA CONSTRUCTION FROM STREAMING VIDEO FOR POWER- CONSTRAINED MOBILE MULTIMEDIA ENVIRONMENTS by XUNYU PAN (Under the Drecton of Suchendra M. Bhandarkar) ABSTRACT In modern tmes, more and more
More informationBrushlet Features for Texture Image Retrieval
DICTA00: Dgtal Image Computng Technques and Applcatons, 1 January 00, Melbourne, Australa 1 Brushlet Features for Texture Image Retreval Chbao Chen and Kap Luk Chan Informaton System Research Lab, School
More informationArtificial Intelligence (AI) methods are concerned with. Artificial Intelligence Techniques for Steam Generator Modelling
Artfcal Intellgence Technques for Steam Generator Modellng Sarah Wrght and Tshldz Marwala Abstract Ths paper nvestgates the use of dfferent Artfcal Intellgence methods to predct the values of several contnuous
More informationAir Transport Demand. Ta-Hui Yang Associate Professor Department of Logistics Management National Kaohsiung First Univ. of Sci. & Tech.
Ar Transport Demand Ta-Hu Yang Assocate Professor Department of Logstcs Management Natonal Kaohsung Frst Unv. of Sc. & Tech. 1 Ar Transport Demand Demand for ar transport between two ctes or two regons
More informationChinese Word Segmentation based on the Improved Particle Swarm Optimization Neural Networks
Chnese Word Segmentaton based on the Improved Partcle Swarm Optmzaton Neural Networks Ja He Computatonal Intellgence Laboratory School of Computer Scence and Engneerng, UESTC Chengdu, Chna Department of
More informationPYTHON IMPLEMENTATION OF VISUAL SECRET SHARING SCHEMES
PYTHON IMPLEMENTATION OF VISUAL SECRET SHARING SCHEMES Ruxandra Olmd Faculty of Mathematcs and Computer Scence, Unversty of Bucharest Emal: ruxandra.olmd@fm.unbuc.ro Abstract Vsual secret sharng schemes
More informationGA-Based Learning Algorithms to Identify Fuzzy Rules for Fuzzy Neural Networks
Seventh Internatonal Conference on Intellgent Systems Desgn and Applcatons GA-Based Learnng Algorthms to Identfy Fuzzy Rules for Fuzzy Neural Networks K Almejall, K Dahal, Member IEEE, and A Hossan, Member
More informationSimulation: Solving Dynamic Models ABE 5646 Week 11 Chapter 2, Spring 2010
Smulaton: Solvng Dynamc Models ABE 5646 Week Chapter 2, Sprng 200 Week Descrpton Readng Materal Mar 5- Mar 9 Evaluatng [Crop] Models Comparng a model wth data - Graphcal, errors - Measures of agreement
More informationProper Choice of Data Used for the Estimation of Datum Transformation Parameters
Proper Choce of Data Used for the Estmaton of Datum Transformaton Parameters Hakan S. KUTOGLU, Turkey Key words: Coordnate systems; transformaton; estmaton, relablty. SUMMARY Advances n technologes and
More informationCorrelative features for the classification of textural images
Correlatve features for the classfcaton of textural mages M A Turkova 1 and A V Gadel 1, 1 Samara Natonal Research Unversty, Moskovskoe Shosse 34, Samara, Russa, 443086 Image Processng Systems Insttute
More informationModule Management Tool in Software Development Organizations
Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,
More informationOPTIMAL CONFIGURATION FOR NODES IN MIXED CELLULAR AND MOBILE AD HOC NETWORK FOR INET
OPTIMAL CONFIGURATION FOR NODE IN MIED CELLULAR AND MOBILE AD HOC NETWORK FOR INET Olusola Babalola D.E. Department of Electrcal and Computer Engneerng Morgan tate Unversty Dr. Rchard Dean Faculty Advsor
More informationSVM-based Learning for Multiple Model Estimation
SVM-based Learnng for Multple Model Estmaton Vladmr Cherkassky and Yunqan Ma Department of Electrcal and Computer Engneerng Unversty of Mnnesota Mnneapols, MN 55455 {cherkass,myq}@ece.umn.edu Abstract:
More informationData Mining: Model Evaluation
Data Mnng: Model Evaluaton Aprl 16, 2013 1 Issues: Evaluatng Classfcaton Methods Accurac classfer accurac: predctng class label predctor accurac: guessng value of predcted attrbutes Speed tme to construct
More informationGeneral Vector Machine. Hong Zhao Department of Physics, Xiamen University
General Vector Machne Hong Zhao (zhaoh@xmu.edu.cn) Department of Physcs, Xamen Unversty The support vector machne (SVM) s an mportant class of learnng machnes for functon approach, pattern recognton, and
More informationIMPACT OF RADIO MAP SIMULATION ON POSITIONING IN INDOOR ENVIRONTMENT USING FINGER PRINTING ALGORITHMS
IMPACT OF RADIO MAP SIMULATION ON POSITIONING IN INDOOR ENVIRONTMENT USING FINGER PRINTING ALGORITHMS Jura Macha and Peter Brda Unversty of Zlna, Faculty of Electrcal Engneerng, Department of Telecommuncatons
More informationA SYSTOLIC APPROACH TO LOOP PARTITIONING AND MAPPING INTO FIXED SIZE DISTRIBUTED MEMORY ARCHITECTURES
A SYSOLIC APPROACH O LOOP PARIIONING AND MAPPING INO FIXED SIZE DISRIBUED MEMORY ARCHIECURES Ioanns Drosts, Nektaros Kozrs, George Papakonstantnou and Panayots sanakas Natonal echncal Unversty of Athens
More informationOutline. Type of Machine Learning. Examples of Application. Unsupervised Learning
Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton
More informationLearning the Kernel Parameters in Kernel Minimum Distance Classifier
Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department
More informationUser Authentication Based On Behavioral Mouse Dynamics Biometrics
User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA
More informationArray transposition in CUDA shared memory
Array transposton n CUDA shared memory Mke Gles February 19, 2014 Abstract Ths short note s nspred by some code wrtten by Jeremy Appleyard for the transposton of data through shared memory. I had some
More informationA Statistical Model Selection Strategy Applied to Neural Networks
A Statstcal Model Selecton Strategy Appled to Neural Networks Joaquín Pzarro Elsa Guerrero Pedro L. Galndo joaqun.pzarro@uca.es elsa.guerrero@uca.es pedro.galndo@uca.es Dpto Lenguajes y Sstemas Informátcos
More informationLobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide
Lobachevsky State Unversty of Nzhn Novgorod Polyhedron Quck Start Gude Nzhn Novgorod 2016 Contents Specfcaton of Polyhedron software... 3 Theoretcal background... 4 1. Interface of Polyhedron... 6 1.1.
More informationAudio Event Detection and classification using extended R-FCN Approach. Kaiwu Wang, Liping Yang, Bin Yang
Audo Event Detecton and classfcaton usng extended R-FCN Approach Kawu Wang, Lpng Yang, Bn Yang Key Laboratory of Optoelectronc Technology and Systems(Chongqng Unversty), Mnstry of Educaton, ChongQng Unversty,
More informationCLASSIFICATION OF ULTRASONIC SIGNALS
The 8 th Internatonal Conference of the Slovenan Socety for Non-Destructve Testng»Applcaton of Contemporary Non-Destructve Testng n Engneerng«September -3, 5, Portorož, Slovena, pp. 7-33 CLASSIFICATION
More information