HYBRID OFF-LINE OCR FOR ISOLATED HANDWRITTEN GREEK CHARACTERS

Size: px
Start display at page:

Download "HYBRID OFF-LINE OCR FOR ISOLATED HANDWRITTEN GREEK CHARACTERS"

Transcription

1 HYBRID OFF-LINE OCR FOR ISOLATED HANDWRITTEN GREEK CHARACTERS G. amvakas,2, B. Gatos, I. ratkaks, N. Stamatopoulos,2, A. Ronots 2, S.J. erantons Computatonal Intellgence Laboratory Insttute of Informatcs and Telecommuncatons Natonal Center for Scentfc Research Demokrtos GR-53 Aga araskev, Athens, Greece {gbam, bgat, pratka, nstam, 2 Department of Informatcs and Telecommuncatons Unversty of Athens, Greece ABSTRACT In ths paper, we present an off-lne OCR methodology for solated handwrtten Greek characters manly based on a robust hybrd feature extracton scheme. Frst, mage pre-processng s performed n order to normalze the character mages as well as to correct character slant. At the next step, two types of features are combned n a hybrd fashon. The frst one dvdes the character mage nto a set of zones and calculates the densty of the character pxels n each zone. In the second type of features, the area that s formed from the projectons of the upper and lower as well as of the left and rght character profles s calculated. For the classfcaton step Support ectors Machnes (SM are used. The performance of the proposed methodology s demonstrated after testng wth the CIL database (handwrtten Greek character database, whch was created from dfferent wrters. KEY WORDS Handwrtten OCR, Hybrd Features, Handwrtten Character Database. Introducton Optcal character recognton (OCR s one of the most successful applcatons of automatc pattern recognton. The OCR stems attempt to facltate the every-day use of computers n the transformaton of large amount of documents, ether prnted or handwrtten, nto electronc form for further processng. Nowada, the recognton of prnted solated characters s performed wth hgh accuracy. However, the recognton of handwrtten characters stll remans an open search problem. In general, character recognton procedure conssts of two steps: (a feature extracton where each character s represented as a feature vector and (b classfcaton of the vectors nto a number of classes []. To develop hgh n at these two ponts. To get the outer profles, for each y value, the outermost x values of each contour half s selected. To get the nner profles, for each y value, the nnermost x values are selected. Horzontal profles can accuracy handwrtten character recognton stems, researchers have taken great efforts searchng for effcent feature representaton and classfers wth good generalzaton ablty. Selecton of a feature extracton method s probable the sngle most mportant factor n achevng hgh recognton performance [2]. Due to the nature of handwrtng wth ts hgh degree of varablty and mprecson, obtanng these features s a dffcult task. A feature extracton algorthm must be robust enough that for a varety of nstances of the same symbol, smlar feature sets are generated, thereby makng the subsequent classfcaton task less dffcult [3]. In the lterature, feature extracton methods have been based on two types of features: statstcal and structural [4]. Representaton of a character mage by statstcal dstrbuton takes care of style varatons to some extent. Ths method s used for reducng the dmenson of the feature set provdng hgh speed and low complexty [5]. On the other hand, characters can be represented by structural features wth hgh tolerance to dstortons and style varatons. The most common statstcal features used for character representaton are: (a zonng, where the character s dvded nto several zones and features are extracted from the denstes n each zone, (b projectons, where the twodmensonal mage (2-D s represented as onedmensonal sgnal (-D [6] and (c crossngs and dstances [5]. For example, contour drecton features measure the drecton of the contour of the character [7], whch are generated by dvdng the mage array nto rectangular and dagonal zones and computng hstograms of chan codes n these zones. In [2], a feature extracton method whch s based on horzontal and vertcal profles of the contour of the characters s presented. Frst the horzontal projecton s computed locatng the uppermost and lowermost pxels of the contour. The contour s splt be extracted n a smlar fashon, startng by dvdng the contour n upper and lower halves. These features, although ndependent to nose and deformaton, depend on rotaton. Another example s presented n [8], where

2 two sets of features are extracted usng crossngs and dstances. The frst one conssts of the number of transtons from background to foreground pxels along vertcal and horzontal lnes through the character mage and the second one calculates the dstances of the frst mage pxel detected from the upper and lower boundares, of the mage, along vertcal lnes and from the left and rght boundares along horzontal lnes. Structural features are based on topologcal and geometrcal propertes of the character, such as maxma and mnma, cusps above and below a threshold, cross ponts, branch ponts, strokes and ther drectons, nflecton between two ponts, horzontal curves at top or bottom, etc. One of the most popular technques for structural feature extracton s codng whch s obtaned by mappng the strokes of a character nto o 2-D parameter space, whch s made up the codes. Another method for structural feature representaton s representaton by graphs. The character s frst parttoned nto a set of topologcal prmtves, such as strokes, loops, cross ponts etc. and, these prmtves are represented usng attrbuted or relatonal graphs [5].There are two knds of mage representaton by graphs. The frst knd uses the coordnates of the character shape [9]. The second one s an abstract representaton wth nodes correspondng to the strokes and edges correspondng to the relatonshps between the strokes []. For the recognton of handwrtten Greek characters the only etng approach s based on structural features []. A 28-dmenson vector s extracted consstng of hstograms and profles. The well known horzontal and vertcal hstograms are used n combnaton wth the radal hstogram, out-n radal and n-out radal profles. Assume that each character s normalzed to a 32x32 matrx. Consder that the value of the element n the m-th row and the n-th column of the character matrx s gven by a functon f: f ( mn, = a ( mn where a mn takes bnary values (.e., for whte pxels and for black pxels. The horzontal hstogram of the H h character matrx s the sum of black pxels n each row (.e., 32 features: H ( (, h m = f m n (2 n Smlarly, the vertcal hstogram H v of the character matrx s the sum of black pxels n each column (.e., 32 features: H ( (, v n = f m n (3 m The radal hstogram s defned as the sum of foreground pxels on a radus that starts from the center of the mage and ends up at the border. The radal hstogram values are calculated wth a step of 5 degrees (.e., 72 features: 6 Hr ( φ = f( 6 sn φ, 6 + cos φ (4 = φ = 5* k, k [,72 The value of the out-n radal profle s defned as the poston of the frst foreground pxel found on the radus that starts from the perphery and goes to the center of the character mage formng an angle φ wth the horzontal a. The out-n radal profle values are calculated wth a step of 5 degrees (.e., 72 features: I I : f( 6 sn φ, 6 + cos φ o ( φ = = 6 (5 & f ( 6 sn φ, 6 cos φ Ι +Ι φ = 5* k, k [,72 Smlarly, the value of the n-out profle s defned as the poston of the frst foreground pxel found on the radus that starts from center of the character mage and goes to the perphery formng an angle φ wth the vertcal a. The n-out radal profle values are calculated wth a step of 5 degrees (.e., 72 features: J J : f( 6 sn φ, 6 + cos φ o ( φ = (6 = 6 & f( 6 Jsn φ, 6 Jcos φ + φ = 5* k, k [,72 Thus, a 28-dmenson vector s extracted for each character, as presented n Fg. : (a (b (c (d (e (f Fg. : a Character Matr b Horzontal Hstogram, c ertcal Hstogram d Radal Hstogram, e Radal out-n rofle, f Radal n-out rofle. In ths paper, we present a hybrd statstcal feature extracton method for solated handwrtten Greek characters, whch s based on a combnaton of features extracted from dvdng the character mage nto zones and the projectons of the upper, lower, left and rght profles. The expermental results ndcate the 98

3 effectveness of our approach. We manly compare our method wth the only etng method for recognzng handwrtten Greek characters, whch s descrbed n []. The remanng of the paper s organzed as follows. In Secton 2 our approach s presented whle expermental results are dscussed n Secton 3. Fnally, conclusons are drawn n Secton roposed Method 2.. re-processng Before the feature extracton algorthm takes place, we frst normalze all bnary character mages to a ΝxN matrx. After normalzaton, slope correcton s performed, whch s based on [2]. The domnant slope of the character s found from the slope corrected character whch gves the mnmum entropy of a vertcal projecton hstogram. The vertcal hstogram projecton s calculated for a range of slope correcton angles, α, where the angle s gven n ± θ. A slope correcton range of θ=6 ο appears to cover all wrtng styles. The slope of the character α m, s found from: α = mn H (7 m H = N = a ± θ log where N s the number of the columns n the vertcal projecton hstogram and s the probablty of a foreground pxel appearng n column. The character s then corrected by α m usng (see Fg. 2: (8 x = x y tan( a m (9 y = y ( (a (b Fg. 2: Feature extracton of a character mage based on zones. (a The normalzed character mage, (b Features based on zones. Darker squares ndcate hgher densty of character pxels. we calculate the densty of the character pxels (see Fg. 2. Let Z H and Z be the total number of zones formed n both horzontal and vertcal drecton. Then, features based on zones f z (, = Z H Z - are calculated as follows: where, x ( y ( e e z f ( = m( ( ( ( x ( = ( ZH ZH Z max, max xe( = ( ZH + Z H H ZH ymax ( =, ymax ( = ( ZH Z + ZH Z In case of features based on character (upper/lower profle projectons, the character mage s dvded nto two sectons separated by the horzontal lne y = y t (see Eq. 2: x ym( y (2 yt = m( x y x 2.2. Feature Extracton In our approach we employ two types of features. The frst one dvdes the character mage nto a set of zones and calculates the densty of the character pxels n each zone. In the second type of features, the area that s formed from the projectons of the upper and lower as well as of the left and rght character profles s calculated. Let m( be the character mage array havng s for foreground and s for background pxels, x max and y max be the wdth and the heght of the character mage. In the case of features based on zones, the mage s dvded nto horzontal and vertcal zones. For each zone, (a (b (c Fg. 3: Feature extracton of a character mage based on upper and lower character profle projectons. (a The normalzed character mage. (b Upper and lower character profles. (c The extracted features. Darker squares ndcate hgher densty of zone pxels. Upper/lower profles (Eq. 3, 4 are computed by consderng, for each mage column, the dstance between the horzontal lne y t and the closest pxel to the upper/lower boundary of the character mage (see Fg. 3. where y where y yup ( x = yt y, (3 yt yt, f m( = = y :( m( = & y = mn(y, y [,yt ],else ylo ( x = y yt, (4 ymax yt, f m( = = yt y:( m( = &y = max(, y [y t,y max ],else 99

4 Let be the total number of blocks formed n each produced zone (upper, lower. For each block, we calculate the area of the upper/lower character profles denoted as n the followng: xe ( up _ ar ( = yup( x ( f ( R le _ ar ( = xle( ( f ( R r _ ar ( = xr( ( f where, (5 ymax, ymax y x ( s( = ( R y e( = ( R + e R R f lo _ ar ( = ylo( x R R (6 ( and =.. R -. Fg. 4 llustrates the features extracted where, from a character mage usng projectons of character profles. xmax ( = (, xmax xe( = ( + The overall calculaton of the proposed feature vector s and =.. -. Fg. 3 llustrates the features extracted gven n Eq. 22. The correspondng feature vector length from a character mage usng projectons of character equals to Z H Z R. profles. x ( y e e( z f ( = mx (,, =... ZHZ In case of features based on character (left/rght profle ( ( projectons, the character mage s dvded nto two xe ( Z HZv f up_ ar( = y up( x, = ZHZ... ZHZ + sectons separated by the vertcal lne x = x t (see Eq. 7 ( Z HZ xe ( Z HZ + (22 m( x f( = f lo_ ar( = y lo( x, = ZHZ ZHZ + 2 x y xt = ( Z HZ + v m( (7 x y ( Z HZ + 2v R f le_ ar( = x le(, = ZHZ ZHZ R ( Z HZ + 2v ( Z HZ R R f r_ ar( = = = x r(, ZHZ 2 R... ZHZ 2 2R ( Z HZ R 3. Expermental Results (2 (2 (a (b (c Fg. 4: Feature extracton of a character mage based on left and rght character profle projectons. (a The normalzed character mage. (b Left and rght character profles; (c The extracted features. Darker squares ndcate hgher densty of zone pxels. Left/rght character profles (Eq. 8,9 are computed by consderng, for each mage column, the dstance between the vertcal lne x t and the closest character pxel to the left/rght boundary of the character mage (see Fg. 4 : where x x ( = x x, le t (8 xt xt, f m( = = x :( m( = & x = mn(x, x [,xt ], else x ( = x x, r t (9 xmax xt, f m( = where x = xt x:( m( = & x = max(x, x [xt,x max ],else Let R be the total number of blocks formed n each produced zone (left, rght. For each block, we calculate the area of left/rght character profles denoted as n the followng: For our experments, we have used the CIL Greek characters database. Ths database was created by handlng a number of forms (see Fg. 5 below that nclude 56 Greek characters. Intally, wrters are requested to fll these forms and afterwards, a scanner converts them to dgtal bnary mages. The last stage deals wth character extracton: Frstly, t detects lne and column ntersectons of each form and thus, approxmates the coordnates of cells. However, very slght nclnaton of forms, due to scannng, may cause serously erroneous cell detecton. Then, n order to avod ths, each cell s coordnates are re-defned, by scannng cell upwards and downwards untl border lnes are found. Lastly, characters are detected wthn cells and are sorted wthn the CIL database. Every wrter flled 5 forms, resultng n a data base of 28, solated and labeled characters. Each class represents a character and conssts of 5 varatons of ths character (56x5=28,. After the sze normalzaton step some characters such as the upper-case Ο and the lower-case o, are consdered to be the same. So, for havng meanngful results we merged these two classes nto one, by randomly selectng 5 characters from both classes. Ths was done to a total of par of classes. Table shows whch classes are merged. Ths concluded n havng 46 classes wth 5 patterns n each class and the database now has 46x5=23, characters 2

5 features based on zones and features based on character profle projectons led to the feature extracton model (Eq. 22 that uses a total of sxty-fve (65 features. In the partcular bnary classfcaton problem classfcaton step was performed usng two well-known classfcaton algorthms, Eucldean Mnmum Dstance Classfer (EMDC [3] and Support ector Machnes (SM [4]. Formally, the support vector machnes (SM requre the soluton of an optmzaton problem, gven a tranng set of nstance-label pars (x, y, =,, m, where n x R and y {, } m. The optmzaton problem s defned as follows: mn ω, b, ξ T ωω+ C 2 T subject to y( ω φ( x + b ξ ξ m ξ = (23 Fg. 5: Sample of the forms used for the CIL Greek Characters Data Base. Moreover, /5 of each class was used for testng and the 4/5 for tranng. So the used database was splt nto a tranng set of 8,4 characters (4x46=8,4 and a testng set of 4,6 characters (x46= 4,6. Table : Merged Classes Upper-case Lower-case Ε ε 2 Θ θ 3 Κ κ 4 Ο ο 5 Π π 6 Ρ ρ 7 Τ τ 8 Φ φ 9 Χ χ Ψ ψ As t has already been descrbed n Sectons 2 we have used a sze normalzaton step followed by a slope correcton step before feature extracton. Durng the normalzaton step, the sze of the normalzed character mages used s x max =6 and y max =6. In the case of features based on zones, the character mage s dvded nto fve (Z H =5 horzontal and fve (Z =5 vertcal zones formng a total of twenty-fve (25 blocks wth sze 2x2 (see Fg. 2. Therefore, the total number of features s 25. In the case of features based on character (upper/lower profle projectons we keep the same sze of the normalzed mage, whle the mage s dvded nto ten ( vertcal zones ( = (see Fg. 3. Consequently, the total number of features equals to twenty (2. Smlarly, the normalzed mage dvded nto ten ( horzontal zones (R = (see Fg. 4. Therefore, the total number of features equals to twenty (2. Combnaton of Regardng the classfcaton step, SM was used n conjuncton wth the Radal Bass Functon (RBF kernel, a popular, general-purpose t powerful kernel, denoted as: 2 K( x, x exp( γ x x (24 j j Furthermore, a grd search was performed n order to fnd the optmal values for both the varance parameter (γ of the RBF kernel and the cost parameter (C of SM (see Eq. 23. Table 2 depcts the recognton rates (% for the proposed hybrd method, the sngle feature methods and the method descrbed n [] whch s the only method reported n the lterature for the recognton of Greek handwrtten characters. These recognton rates acheved after combnng slope correcton wth ether sngle features of both feature extracton schemes. We can draw several conclusons. Frst, as t can been easly seen the best performance, for both classfers, s acheved n the case of usng an addtve fuson resulted after the combnaton of slope correcton precedng the hybrd feature extracton scheme. Moreover, our approach seems to have better results although the total number of features used (65 s smaller than the one n [] (28. Fnally, the SM acheved consderably hgher recognton rates than the EMDC. 4. Concluson Ths paper proposes an off-lne OCR stem for solated handwrtten Greek characters that s based on an addtve fuson resulted after a novel combnaton of two dfferent modes of character mage normalzaton (sze normalzaton and slope correcton and robust hybrd feature extracton. 2

6 We compared the proposed approach wth the method descrbed n [], whch s the only work n recognzng Greek handwrtten characters. Our approach acheved hgher rates of recognton. Our future research wll focus on explotng new features as well as fuson methods to further mprove the current performance. Acknowledgements Ths research s carred out wthn the framework of the Greek Mnstry of Research funded R&D project OLYTIMO [5] whch ams to process and provde access to the content of valuable hstorcal books and handwrtten manuscrpts. References [] S. Brtto, R. Sabourn, F. Bortolozz, C. Y. Suen, Foreground and Backround Informaton n an HMM- Based Method for Recognton of Isolated Characters and Numeral Strngs, 9th Internatonal Workshop on Fronters n Handwrtng Recognton (IWFHR-9, October 26-29, 24, Kokubunj, Tokyo, Japan, pp [2] O. D. Trer, A. K. Jan, T.Taxt, Features Extracton Methods for Character Recognton A Survey, attern Recognton, 29(4 : , 996. [3] J. A. Ftzgerald, F. Geselbrechtnger, and T. Kechad, Applcaton of Fuzzy Logc to Onlne Recognton of Handwrtten Symbols, 9th Internatonal Workshop on Fronters n Handwrtng Recognton (IWFHR 9, Tokyo, Japan, pp , October 26-29, 24. [4].K. Covndan, A.. Shvaprasad, Character Recognton A Revew, attern Recognton, 23(7, 99, pp [5] N. Arca and F. Yarman-ural, An Overvew of Character Recognton Focused on Off-lne Handwrtng, Table 2: Expermental Results re-processng Features Number of Classfer Recognton Slope Correcton Method n Hybrd features EMDC SM Rate (% [] Zones rojectons % % % % % % % % % % % % % % % % IEEE Transactons on Stems, Man, and Cybernetcs, art C: Applcatons and Revews, 3(2, 2, pp [6] Y. Tao, Y. Y. Tang, The feature extracton of Chnese character based on contour nformaton, n roc. 5thInt. Conf. Document Anal. Reognt., Bangalore, Inda, 999, pp [7] K. M. Mohuddn and J. Mao, A Comprehensve Study of Dfferent Classfers for Handprnted Character Recognton. attern Recognton, ractce I, pp , 994. [8] J. H. Km, K. K. Km, C. Y. Suen, Hybrd Schemes Of Homogeneous and Heterogeneous Classfers for Cursve Word Recognton, 7 th Internatonal Workshop on Fronters n Handwrtng Recognton, September -3 2, Amsterdam, ISBN , Njmegen: Internatonal Unpen Foundaton, pp s. [9] H. Cheng, W. H. Hsu, M. C. Kuo, Recognton of handprnted Chnese characters va stroke relaxaton, attern Recognton, 26(4, pp , 993. [] W. Lu, Y. Ren, C. Y. Suen, Herarchcal attrbuted graph representaton and recognton of handwrtten Chnese characters, attern Recognton, 24(7 pp , 99. [] E. Kavalleratou, N. Fotaks, G Kokknaks, Handwrtten Character Recognton Based on Structural Characterstcs, cpr, p. 339, 6 th Internatonal Conference on attern Recognton (ICR'2 - olume 3, 22. [2] R. Buse, Z.Q. Lu, and T. Caell, A Structural and Relatonal Approach to Handwrtten Word Recognton, IEEE Trans. Stems, Man, and Cybernetcs, art B, 27(5, pp , Oct [3] Theodords, S., and Koutroumbas, K., attern Recognton,(Academc ress, 997. [4] Cortes C., and apnk,., Support-vector network, Machne Learnng, vol. 2, pp , 997. [5] OLYTIMO project,

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram Shape Representaton Robust to the Sketchng Order Usng Dstance Map and Drecton Hstogram Department of Computer Scence Yonse Unversty Kwon Yun CONTENTS Revew Topc Proposed Method System Overvew Sketch Normalzaton

More information

A Binarization Algorithm specialized on Document Images and Photos

A 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 information

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content 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 information

Foreground and Background Information in an HMM-based Method for Recognition of Isolated Characters and Numeral Strings

Foreground and Background Information in an HMM-based Method for Recognition of Isolated Characters and Numeral Strings Foreground and Background Informaton n an HMM-based Method for Recognton of Isolated Characters and Numeral Strngs Alceu de S. Brtto Jr a,b, Robert Sabourn c,d, Flavo Bortolozz a,chng Y. Suen d a Pontfíca

More information

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur

FEATURE 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 information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A 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 information

Recognition of Handwritten Numerals Using a Combined Classifier with Hybrid Features

Recognition of Handwritten Numerals Using a Combined Classifier with Hybrid Features Recognton of Handwrtten Numerals Usng a Combned Classfer wth Hybrd Features Kyoung Mn Km 1,4, Joong Jo Park 2, Young G Song 3, In Cheol Km 1, and Chng Y. Suen 1 1 Centre for Pattern Recognton and Machne

More information

Edge Detection in Noisy Images Using the Support Vector Machines

Edge 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 information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism 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 information

Efficient Off-Line Cursive Handwriting Word Recognition

Efficient Off-Line Cursive Handwriting Word Recognition Effcent Off-Lne Cursve Handwrtng Word Recognton B. Gatos, I. Pratkaks, A.L. Kesds, S.J. Perantons To cte ths verson: B. Gatos, I. Pratkaks, A.L. Kesds, S.J. Perantons. Effcent Off-Lne Cursve Handwrtng

More information

Classifier Selection Based on Data Complexity Measures *

Classifier 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 information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A 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 information

Feature Reduction and Selection

Feature 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 information

A high precision collaborative vision measurement of gear chamfering profile

A high precision collaborative vision measurement of gear chamfering profile Internatonal Conference on Advances n Mechancal Engneerng and Industral Informatcs (AMEII 05) A hgh precson collaboratve vson measurement of gear chamferng profle Conglng Zhou, a, Zengpu Xu, b, Chunmng

More information

Local Quaternary Patterns and Feature Local Quaternary Patterns

Local Quaternary Patterns and Feature Local Quaternary Patterns Local Quaternary Patterns and Feature Local Quaternary Patterns Jayu Gu and Chengjun Lu The Department of Computer Scence, New Jersey Insttute of Technology, Newark, NJ 0102, USA Abstract - Ths paper presents

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. 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 information

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1

Outline. 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 information

Support Vector Machines

Support 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 information

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and

More information

Face Recognition University at Buffalo CSE666 Lecture Slides Resources:

Face Recognition University at Buffalo CSE666 Lecture Slides Resources: Face Recognton Unversty at Buffalo CSE666 Lecture Sldes Resources: http://www.face-rec.org/algorthms/ Overvew of face recognton algorthms Correlaton - Pxel based correspondence between two face mages Structural

More information

th International Conference on Document Analysis and Recognition

th International Conference on Document Analysis and Recognition 2013 12th Internatonal Conference on Document Analyss and Recognton Onlne Handwrtten Cursve Word Recognton Usng Segmentaton-free n Combnaton wth P2DBM-MQDF Blan Zhu 1, Art Shvram 2, Srrangaraj Setlur 2,

More information

Using Fuzzy Logic to Enhance the Large Size Remote Sensing Images

Using 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 information

12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification

12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification Introducton to Artfcal Intellgence V22.0472-001 Fall 2009 Lecture 24: Nearest-Neghbors & Support Vector Machnes Rob Fergus Dept of Computer Scence, Courant Insttute, NYU Sldes from Danel Yeung, John DeNero

More information

A New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1

A New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1 A New Feature of Unformty of Image Texture Drectons Concdng wth the Human Eyes Percepton Xng-Jan He, De-Shuang Huang, Yue Zhang, Tat-Mng Lo 2, and Mchael R. Lyu 3 Intellgent Computng Lab, Insttute of Intellgent

More information

Machine Learning: Algorithms and Applications

Machine 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 information

Detection of an Object by using Principal Component Analysis

Detection 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 information

Support Vector Machines

Support Vector Machines Support Vector Machnes Decson surface s a hyperplane (lne n 2D) n feature space (smlar to the Perceptron) Arguably, the most mportant recent dscovery n machne learnng In a nutshell: map the data to a predetermned

More information

An Improved Image Segmentation Algorithm Based on the Otsu Method

An Improved Image Segmentation Algorithm Based on the Otsu Method 3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,

More information

Positive Semi-definite Programming Localization in Wireless Sensor Networks

Positive 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 information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning 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 information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement 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 information

A Clustering Algorithm for Key Frame Extraction Based on Density Peak

A Clustering Algorithm for Key Frame Extraction Based on Density Peak Journal of Computer and Communcatons, 2018, 6, 118-128 http://www.scrp.org/ournal/cc ISSN Onlne: 2327-5227 ISSN Prnt: 2327-5219 A Clusterng Algorthm for Key Frame Extracton Based on Densty Peak Hong Zhao

More information

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements

2x 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 information

Unsupervised Learning

Unsupervised Learning Pattern Recognton Lecture 8 Outlne Introducton Unsupervsed Learnng Parametrc VS Non-Parametrc Approach Mxture of Denstes Maxmum-Lkelhood Estmates Clusterng Prof. Danel Yeung School of Computer Scence and

More information

Related-Mode Attacks on CTR Encryption Mode

Related-Mode Attacks on CTR Encryption Mode Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory

More information

Invariant Shape Object Recognition Using B-Spline, Cardinal Spline, and Genetic Algorithm

Invariant Shape Object Recognition Using B-Spline, Cardinal Spline, and Genetic Algorithm Proceedngs of the 5th WSEAS Int. Conf. on Sgnal Processng, Robotcs and Automaton, Madrd, Span, February 5-7, 6 (pp4-45) Invarant Shape Obect Recognton Usng B-Splne, Cardnal Splne, and Genetc Algorthm PISIT

More information

Face Recognition Based on SVM and 2DPCA

Face Recognition Based on SVM and 2DPCA Vol. 4, o. 3, September, 2011 Face Recognton Based on SVM and 2DPCA Tha Hoang Le, Len Bu Faculty of Informaton Technology, HCMC Unversty of Scence Faculty of Informaton Scences and Engneerng, Unversty

More information

An Image Fusion Approach Based on Segmentation Region

An 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 information

A Gradient Difference based Technique for Video Text Detection

A Gradient Difference based Technique for Video Text Detection A Gradent Dfference based Technque for Vdeo Text Detecton Palaahnakote Shvakumara, Trung Quy Phan and Chew Lm Tan School of Computng, Natonal Unversty of Sngapore {shva, phanquyt, tancl }@comp.nus.edu.sg

More information

PRÉSENTATIONS DE PROJETS

PRÉSENTATIONS DE PROJETS PRÉSENTATIONS DE PROJETS Rex Onlne (V. Atanasu) What s Rex? Rex s an onlne browser for collectons of wrtten documents [1]. Asde ths core functon t has however many other applcatons that make t nterestng

More information

A Gradient Difference based Technique for Video Text Detection

A Gradient Difference based Technique for Video Text Detection 2009 10th Internatonal Conference on Document Analyss and Recognton A Gradent Dfference based Technque for Vdeo Text Detecton Palaahnakote Shvakumara, Trung Quy Phan and Chew Lm Tan School of Computng,

More information

The Study of Remote Sensing Image Classification Based on Support Vector Machine

The Study of Remote Sensing Image Classification Based on Support Vector Machine Sensors & Transducers 03 by IFSA http://www.sensorsportal.com The Study of Remote Sensng Image Classfcaton Based on Support Vector Machne, ZHANG Jan-Hua Key Research Insttute of Yellow Rver Cvlzaton and

More information

Lecture #15 Lecture Notes

Lecture #15 Lecture Notes Lecture #15 Lecture Notes The ocean water column s very much a 3-D spatal entt and we need to represent that structure n an economcal way to deal wth t n calculatons. We wll dscuss one way to do so, emprcal

More information

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan

More information

The Research of Ellipse Parameter Fitting Algorithm of Ultrasonic Imaging Logging in the Casing Hole

The Research of Ellipse Parameter Fitting Algorithm of Ultrasonic Imaging Logging in the Casing Hole Appled Mathematcs, 04, 5, 37-3 Publshed Onlne May 04 n ScRes. http://www.scrp.org/journal/am http://dx.do.org/0.436/am.04.584 The Research of Ellpse Parameter Fttng Algorthm of Ultrasonc Imagng Loggng

More information

Optimal Design of Nonlinear Fuzzy Model by Means of Independent Fuzzy Scatter Partition

Optimal Design of Nonlinear Fuzzy Model by Means of Independent Fuzzy Scatter Partition Optmal Desgn of onlnear Fuzzy Model by Means of Independent Fuzzy Scatter Partton Keon-Jun Park, Hyung-Kl Kang and Yong-Kab Km *, Department of Informaton and Communcaton Engneerng, Wonkwang Unversty,

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

Accounting for the Use of Different Length Scale Factors in x, y and z Directions

Accounting for the Use of Different Length Scale Factors in x, y and z Directions 1 Accountng for the Use of Dfferent Length Scale Factors n x, y and z Drectons Taha Soch (taha.soch@kcl.ac.uk) Imagng Scences & Bomedcal Engneerng, Kng s College London, The Rayne Insttute, St Thomas Hosptal,

More information

Discriminative Dictionary Learning with Pairwise Constraints

Discriminative Dictionary Learning with Pairwise Constraints Dscrmnatve Dctonary Learnng wth Parwse Constrants Humn Guo Zhuoln Jang LARRY S. DAVIS UNIVERSITY OF MARYLAND Nov. 6 th, Outlne Introducton/motvaton Dctonary Learnng Dscrmnatve Dctonary Learnng wth Parwse

More information

Research and Application of Fingerprint Recognition Based on MATLAB

Research and Application of Fingerprint Recognition Based on MATLAB Send Orders for Reprnts to reprnts@benthamscence.ae The Open Automaton and Control Systems Journal, 205, 7, 07-07 Open Access Research and Applcaton of Fngerprnt Recognton Based on MATLAB Nng Lu* Department

More information

BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET

BOOSTING 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 information

Histogram of Template for Pedestrian Detection

Histogram of Template for Pedestrian Detection PAPER IEICE TRANS. FUNDAMENTALS/COMMUN./ELECTRON./INF. & SYST., VOL. E85-A/B/C/D, No. xx JANUARY 20xx Hstogram of Template for Pedestran Detecton Shaopeng Tang, Non Member, Satosh Goto Fellow Summary In

More information

Classification / Regression Support Vector Machines

Classification / 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 information

Image Representation & Visualization Basic Imaging Algorithms Shape Representation and Analysis. outline

Image 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 information

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

Machine Learning. Topic 6: Clustering

Machine Learning. Topic 6: Clustering Machne Learnng Topc 6: lusterng lusterng Groupng data nto (hopefully useful) sets. Thngs on the left Thngs on the rght Applcatons of lusterng Hypothess Generaton lusters mght suggest natural groups. Hypothess

More information

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION

A 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 information

Incremental MQDF Learning for Writer Adaptive Handwriting Recognition 1

Incremental MQDF Learning for Writer Adaptive Handwriting Recognition 1 200 2th Internatonal Conference on Fronters n Handwrtng Recognton Incremental MQDF Learnng for Wrter Adaptve Handwrtng Recognton Ka Dng, Lanwen Jn * School of Electronc and Informaton Engneerng, South

More information

Face Recognition using 3D Directional Corner Points

Face Recognition using 3D Directional Corner Points 2014 22nd Internatonal Conference on Pattern Recognton Face Recognton usng 3D Drectonal Corner Ponts Xun Yu, Yongsheng Gao School of Engneerng Grffth Unversty Nathan, QLD, Australa xun.yu@grffthun.edu.au,

More information

Parallel matrix-vector multiplication

Parallel matrix-vector multiplication Appendx A Parallel matrx-vector multplcaton The reduced transton matrx of the three-dmensonal cage model for gel electrophoress, descrbed n secton 3.2, becomes excessvely large for polymer lengths more

More information

WATERSHED ALGORITHM BASED SEGMENTATION FOR HANDWRITTEN TEXT IDENTIFICATION

WATERSHED ALGORITHM BASED SEGMENTATION FOR HANDWRITTEN TEXT IDENTIFICATION ISSN: 976-92(ONLINE) DOI:.297/jvp.24. ICTACT JOURNAL ON IMAGE AND VIDEO PROCESSING, FEBRUARY 24, VOLUME: 4, ISSUE: 3 WATERSHED ALGORITHM BASED SEGMENTATION FOR HANDWRITTEN TEXT IDENTIFICATION P. Mathvanan,

More information

A Deflected Grid-based Algorithm for Clustering Analysis

A Deflected Grid-based Algorithm for Clustering Analysis A Deflected Grd-based Algorthm for Clusterng Analyss NANCY P. LIN, CHUNG-I CHANG, HAO-EN CHUEH, HUNG-JEN CHEN, WEI-HUA HAO Department of Computer Scence and Informaton Engneerng Tamkang Unversty 5 Yng-chuan

More information

Cluster Analysis of Electrical Behavior

Cluster 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 information

The Comparison of Calibration Method of Binocular Stereo Vision System Ke Zhang a *, Zhao Gao b

The 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 information

Machine Learning 9. week

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 information

Face Recognition by Fusing Binary Edge Feature and Second-order Mutual Information

Face Recognition by Fusing Binary Edge Feature and Second-order Mutual Information Face Recognton by Fusng Bnary Edge Feature and Second-order Mutual Informaton Jatao Song, Bejng Chen, We Wang, Xaobo Ren School of Electronc and Informaton Engneerng, Nngbo Unversty of Technology Nngbo,

More information

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines A Modfed Medan Flter for the Removal of Impulse Nose Based on the Support Vector Machnes H. GOMEZ-MORENO, S. MALDONADO-BASCON, F. LOPEZ-FERRERAS, M. UTRILLA- MANSO AND P. GIL-JIMENEZ Departamento de Teoría

More information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace 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 information

An efficient method to build panoramic image mosaics

An efficient method to build panoramic image mosaics An effcent method to buld panoramc mage mosacs Pattern Recognton Letters vol. 4 003 Dae-Hyun Km Yong-In Yoon Jong-Soo Cho School of Electrcal Engneerng and Computer Scence Kyungpook Natonal Unv. Abstract

More information

An Image Compression Algorithm based on Wavelet Transform and LZW

An Image Compression Algorithm based on Wavelet Transform and LZW An Image Compresson Algorthm based on Wavelet Transform and LZW Png Luo a, Janyong Yu b School of Chongqng Unversty of Posts and Telecommuncatons, Chongqng, 400065, Chna Abstract a cylpng@63.com, b y27769864@sna.cn

More information

Machine Learning. Support Vector Machines. (contains material adapted from talks by Constantin F. Aliferis & Ioannis Tsamardinos, and Martin Law)

Machine Learning. Support Vector Machines. (contains material adapted from talks by Constantin F. Aliferis & Ioannis Tsamardinos, and Martin Law) Machne Learnng Support Vector Machnes (contans materal adapted from talks by Constantn F. Alfers & Ioanns Tsamardnos, and Martn Law) Bryan Pardo, Machne Learnng: EECS 349 Fall 2014 Support Vector Machnes

More information

Shape-adaptive DCT and Its Application in Region-based Image Coding

Shape-adaptive DCT and Its Application in Region-based Image Coding Internatonal Journal of Sgnal Processng, Image Processng and Pattern Recognton, pp.99-108 http://dx.do.org/10.14257/sp.2014.7.1.10 Shape-adaptve DCT and Its Applcaton n Regon-based Image Codng Yamn Zheng,

More information

Problem Set 3 Solutions

Problem Set 3 Solutions Introducton to Algorthms October 4, 2002 Massachusetts Insttute of Technology 6046J/18410J Professors Erk Demane and Shaf Goldwasser Handout 14 Problem Set 3 Solutons (Exercses were not to be turned n,

More information

Accumulated-Recognition-Rate Normalization for Combining Multiple On/Off-Line Japanese Character Classifiers Tested on a Large Database

Accumulated-Recognition-Rate Normalization for Combining Multiple On/Off-Line Japanese Character Classifiers Tested on a Large Database 4 th Internatonal Workshop on Multple Classfer Systems (MCS23) Guldford, UK Accumulated-Recognton-Rate Normalzaton for Combnng Multple On/Off-Lne Japanese Character Classfers Tested on a Large Database

More information

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures A Novel Adaptve Descrptor Algorthm for Ternary Pattern Textures Fahuan Hu 1,2, Guopng Lu 1 *, Zengwen Dong 1 1.School of Mechancal & Electrcal Engneerng, Nanchang Unversty, Nanchang, 330031, Chna; 2. School

More information

Computer Animation and Visualisation. Lecture 4. Rigging / Skinning

Computer Animation and Visualisation. Lecture 4. Rigging / Skinning Computer Anmaton and Vsualsaton Lecture 4. Rggng / Sknnng Taku Komura Overvew Sknnng / Rggng Background knowledge Lnear Blendng How to decde weghts? Example-based Method Anatomcal models Sknnng Assume

More information

Gender Classification using Interlaced Derivative Patterns

Gender Classification using Interlaced Derivative Patterns Gender Classfcaton usng Interlaced Dervatve Patterns Author Shobernejad, Ameneh, Gao, Yongsheng Publshed 2 Conference Ttle Proceedngs of the 2th Internatonal Conference on Pattern Recognton (ICPR 2) DOI

More information

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:

Outline. 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 information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

PCA Based Gait Segmentation

PCA Based Gait Segmentation Honggu L, Cupng Sh & Xngguo L PCA Based Gat Segmentaton PCA Based Gat Segmentaton Honggu L, Cupng Sh, and Xngguo L 2 Electronc Department, Physcs College, Yangzhou Unversty, 225002 Yangzhou, Chna 2 Department

More information

IMAGE FUSION TECHNIQUES

IMAGE FUSION TECHNIQUES Int. J. Chem. Sc.: 14(S3), 2016, 812-816 ISSN 0972-768X www.sadgurupublcatons.com IMAGE FUSION TECHNIQUES A Short Note P. SUBRAMANIAN *, M. SOWNDARIYA, S. SWATHI and SAINTA MONICA ECE Department, Aarupada

More information

Fuzzy C-Means Initialized by Fixed Threshold Clustering for Improving Image Retrieval

Fuzzy C-Means Initialized by Fixed Threshold Clustering for Improving Image Retrieval Fuzzy -Means Intalzed by Fxed Threshold lusterng for Improvng Image Retreval NAWARA HANSIRI, SIRIPORN SUPRATID,HOM KIMPAN 3 Faculty of Informaton Technology Rangst Unversty Muang-Ake, Paholyotn Road, Patumtan,

More information

3D vector computer graphics

3D 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 information

The Research of Support Vector Machine in Agricultural Data Classification

The Research of Support Vector Machine in Agricultural Data Classification The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou

More information

Experimental Analysis on Character Recognition using Singular Value Decomposition and Random Projection

Experimental Analysis on Character Recognition using Singular Value Decomposition and Random Projection Expermental Analyss on Character Recognton usng Sngular Value Decomposton and Random Projecton Manjusha K. 1, Anand Kumar M. 2, Soman K. P. 3 Centre for Excellence n Computatonal Engneerng and Networkng,

More information

Feature Extractions for Iris Recognition

Feature Extractions for Iris Recognition Feature Extractons for Irs Recognton Jnwook Go, Jan Jang, Yllbyung Lee, and Chulhee Lee Department of Electrcal and Electronc Engneerng, Yonse Unversty 134 Shnchon-Dong, Seodaemoon-Gu, Seoul, KOREA Emal:

More information

R s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes

R 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 information

An Empirical Comparative Study of Online Handwriting Chinese Character Recognition:Simplified v.s.traditional

An Empirical Comparative Study of Online Handwriting Chinese Character Recognition:Simplified v.s.traditional 2013 12th Internatonal Conference on Document Analyss and Recognton An Emprcal Comparatve Study of Onlne Handwrtng Chnese Recognton:Smplfed v.s.tradtonal Yan Gao, Lanwen Jn +, Wexn Yang School of Electronc

More information

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity ISSN(Onlne): 2320-9801 ISSN (Prnt): 2320-9798 Internatonal Journal of Innovatve Research n Computer and Communcaton Engneerng (An ISO 3297: 2007 Certfed Organzaton) Vol.2, Specal Issue 1, March 2014 Proceedngs

More information

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS P.G. Demdov Yaroslavl State Unversty Anatoly Ntn, Vladmr Khryashchev, Olga Stepanova, Igor Kostern EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS Yaroslavl, 2015 Eye

More information

A Robust Method for Estimating the Fundamental Matrix

A Robust Method for Estimating the Fundamental Matrix Proc. VIIth Dgtal Image Computng: Technques and Applcatons, Sun C., Talbot H., Ourseln S. and Adraansen T. (Eds.), 0- Dec. 003, Sydney A Robust Method for Estmatng the Fundamental Matrx C.L. Feng and Y.S.

More information

Unsupervised Learning and Clustering

Unsupervised Learning and Clustering Unsupervsed Learnng and Clusterng Why consder unlabeled samples?. Collectng and labelng large set of samples s costly Gettng recorded speech s free, labelng s tme consumng 2. Classfer could be desgned

More information

Multi-stable Perception. Necker Cube

Multi-stable Perception. Necker Cube Mult-stable Percepton Necker Cube Spnnng dancer lluson, Nobuuk Kaahara Fttng and Algnment Computer Vson Szelsk 6.1 James Has Acknowledgment: Man sldes from Derek Hoem, Lana Lazebnk, and Grauman&Lebe 2008

More information

Dynamic wetting property investigation of AFM tips in micro/nanoscale

Dynamic wetting property investigation of AFM tips in micro/nanoscale Dynamc wettng property nvestgaton of AFM tps n mcro/nanoscale The wettng propertes of AFM probe tps are of concern n AFM tp related force measurement, fabrcaton, and manpulaton technques, such as dp-pen

More information

Scale Selective Extended Local Binary Pattern For Texture Classification

Scale Selective Extended Local Binary Pattern For Texture Classification Scale Selectve Extended Local Bnary Pattern For Texture Classfcaton Yutng Hu, Zhlng Long, and Ghassan AlRegb Multmeda & Sensors Lab (MSL) Georga Insttute of Technology 03/09/017 Outlne Texture Representaton

More information

Classification of Face Images Based on Gender using Dimensionality Reduction Techniques and SVM

Classification of Face Images Based on Gender using Dimensionality Reduction Techniques and SVM Classfcaton of Face Images Based on Gender usng Dmensonalty Reducton Technques and SVM Fahm Mannan 260 266 294 School of Computer Scence McGll Unversty Abstract Ths report presents gender classfcaton based

More information

An Approach to Real-Time Recognition of Chinese Handwritten Sentences

An Approach to Real-Time Recognition of Chinese Handwritten Sentences An Approach to Real-Tme Recognton of Chnese Handwrtten Sentences Da-Han Wang, Cheng-Ln Lu Natonal Laboratory of Pattern Recognton, Insttute of Automaton of Chnese Academy of Scences, Bejng 100190, P.R.

More information

Fast Computation of Shortest Path for Visiting Segments in the Plane

Fast Computation of Shortest Path for Visiting Segments in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 4 The Open Cybernetcs & Systemcs Journal, 04, 8, 4-9 Open Access Fast Computaton of Shortest Path for Vstng Segments n the Plane Ljuan Wang,, Bo Jang

More information