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

Size: px
Start display at page:

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

Transcription

1 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 Unversdade Católca do Paraná (PUC-PR), R. Imaculada Conceção, Curtba (PR) Brazl b Unversdade Estadual de Ponta Grossa (UEPG), Pr. Santos Andrade S/N, Centro - Ponta Grossa (PR) Brazl c École de Technologe Supéreure (ETS), 1100 Rue Notre Dame Ouest - Montreal (QC) H3C 1K3 - Canada d Centre for Pattern Recognton and Machne Intellgence (CENPARMI), 1455 de Masonneuve Blvd. West, Sute GM Montreal (QC) H3G 1M8 - Canada Abstract In ths paper we combne complementary features based on foreground and background nformaton n an HMM-based classfer to recognze handwrtten solated characters and numeral strngs. A zonng scheme based on column and row models provdes a way of dvdng the character nto zones wthout makng the features sze varant. Ths strategy allows us to avod the character normalzaton, whle t provdes a way of havng nformaton from specfc zones of the character. The expermental results on 10 dgt classes, 52 character classes and 6 classes of numeral strngs of dfferent lengths have shown that the proposed features are hghly dscrmmnant. 1. Introducton Many approaches to solvng the handwrtten character recognton problem have been proposed n recent years due to numerous possble applcatons. Drawng up a taxonomy of these approaches s dffcult, snce ther methodologes overlap. However, research n ths feld has bascally consdered nvestgatng: a) feature extracton methods; b) classfcaton methods; and c) system archtectures based on dfferent strateges, such as combnatons of multple classfers, the use of multple templates, and the use of verfcaton modules. The nvestgaton of feature extracton methods has ganed consderable attenton snce a dscrmnatve feature set s consdered the most mportant factor n achevng hgh recognton performance. In [1] a survey of feature extracton methods for off-lne recognton of segmented characters s presented. The authors descrbe mportant aspects that must be consdered before selectng a specfc feature extracton method. In general, the feature extracton methods for numeral recognton reported n the lterature have been based on two types of features: statstcal and structural. The statstcal features are derved from statstcal dstrbutons of ponts, such as zonng, moments, projecton hstograms or drecton hstograms [2,3]. Structural features are based on topologcal and geometrcal propertes of the character, lke strokes and ther drectons, end-ponts, or ntersectons of segments and loops [4,5]. Many researchers have explored the ntegraton of structural and statstcal nformaton to hghlght dfferent character propertes, snce these types of features are consdered to be complementary. In [5], structural and statstcal nformaton s ntegrated nto a classfer based on Hdden Markov Model (HMM). The authors use stateduraton adapted transton probablty dstrbuton and macro-states to overcome the weakness of the HMMs n modelng structural features. The recognton rate s 96.16% n 2,711 dgt samples extracted from the CEDAR database. Another multfeature-based system s proposed n [6]. In ths work, a combnaton of seven dfferent famles of features s proposed n order to arrve at a complete character descrpton. These features are dvded nto global features (nvarant moments, projectons and profles) and local features (ntersectons wth straght lnes, holes and concave arcs, extremtes, end-ponts and junctons). A set of 53,324 dgts extracted from the NIST database s used to test the system. The recognton, rejecton and substtuton rates are 90.82%, 8.93% and 0.25% respectvely. In [7], a MLP-based classfer based on concavty features acheved a recognton rate of over 99.13% n 60,089 samples of handwrtten dgts of the NIST SD19 database. In ths paper, we combne features extracted from the foreground and the background of character mages n an HMM-based system. The challenge s to acheve recognton rates close to that presented n [7] wth MLPs, however, usng an HMM-based system. HMM have been used to model solated characters, whch are appled to recognze words or numeral strngs. The reason s that HMM can model specfc handwrtng knowledge related to the nteracton between adjacent characters n words or numeral strngs easer than MLPs. Moreover, HMM has been successfully used to provde mplct segmentatonbased methods to recognze words and numeral strngs as

2 n [8]. Wth such an approach t s possble to avod a pror segmentaton of the strng or word nto characters. 2. Feature extracton method The extracton method conssts of scannng the character mage from left-to-rght (column-based features) and from bottom-to-top (row-based features). Foreground and background nformaton are combned n a vector of 47 features: 34 foreground plus 13 background features. 2.1 Foreground features (FF) The FF vector conssts of local and global features calculated takng nto account the foreground pxels of the mage columns or rows. The local features are based on transtons from background to foreground pxels and vce versa. For each transton, the mean drecton and correspondng varance are obtaned by means of statstc estmators. These estmators are more sutable for drectonal observatons, snce they are based on a crcular scale. For nstance, gven the drectonal observatons α 1 = 1 o and α 2 = 359 o, they provde a mean drecton ( α ) of 0 o nstead of 180 o calculated by conventonal estmators. Let α1,..., α,..., α N be a set of drectonal F and sze N. observatons wth dstrbuton ( ) α vectors ( OP,...,,..., ). 1 OP OPN the N coordnates ( cos( α ), sn( α ) ) s defned as: 1 N C = cos N = 1 The center of gravty ( S ) ( α ) or S = sn( α ) 1 N N = 1 C, of These coordnates are used to estmate the mean sze of R, as: 2 2 R = ( C + S ) (3) Then, the crcular mean drecton can be obtaned by solvng one of the followng equatons: C cos ( α ) =, ( α ) R (2) S sn = (4) R Fnally, the crcular varance of α s calculated as: Sα = 1 R 0 S α 1 To estmate α and S α for each transton of a numeral mage, we have consdered { 0, 45, 90, 135, 180, 225, 270, 315 } as the set of F s computed by drectonal observatons, whle ( ) α countng the number of successve black pxels over the drecton α from a transton untl the encounter of a whte pxel. In Fgure 2 the transtons n a column of numeral 5 are enumerated from 1 to 6, and the possble drectonal observatons from transtons 3 and 6 are shown. (5) Fgure 1. Crcular mean drecton α and varance S α for a dstrbuton F ( α ) Fgure 1 shows that α represents the angle between the unt vector OP and the horzontal axs, whle P s the ntersecton pont between OP and the unt crcle. The cartesan coordnates of P are defned as: ( ( ), sn( )) cos α α (1) The crcular mean drecton α of the N drectonal observatons on the unt crcle corresponds to the drecton of the resultng vector ( R ) obtaned by the sum of the unt Fgure 2. Transtons n a column mage of numeral 5, and the drectonal observatons to estmate the mean drecton for transtons 3 and 6 In addton to ths drectonal nformaton, we have calculated two other local features: a) relatve poston of each transton, takng nto account the top of the dgt boundng box, and b) whether the transton belongs to the outer or nner contour, whch shows the presence of loops n the numeral mage. Snce for each column we consder 8 possble transtons, at ths pont our feature vector s composed of 32 features.

3 The global features are based on vertcal projecton (VP) of black pxels for each column, and the dervatve of VP between adjacent columns. Ths consttutes a total of 34 features normalzed between 0 and Background features (BF) The BF vector s based on concavty nformaton. These features are used to hghlght the topologcal and geometrcal propertes of the character classes. Each concavty feature represents the number of whte pxels that belong to a specfc concavty confguraton. The label for each whte pxel s chosen based on the Freeman code wth four drectons. Each drecton s explored untl the encounter of a black pxel or the lmts mposed by the dgt boundng box. A whte pxel s labeled f at least two consecutve drectons fnd black pxels. Thus, we have 9 possble concavty confguratons. Moreover, we consder four more confguratons, n order to detect more precsely the presence of loops. The total length of ths feature vector s then 13. The concavty vector s normalzed between 0 and 1, by the total of the concavty codes computed for each column or row of the character mage. Fgure 3 shows the 9 concavty confguratons and also 4 confguratons (A,B,C,D) for false loops A B C D - a black pxel was found n ths drecton - no black pxel was found n ths drecton Fgure 3 Concavty confguratons 2.3 Column and row-based features The feature vector composed of foreground and background features s extracted from each column and row of the character mage. Each feature vector s mapped to one of 256 possble dscrete symbols avalable n a 9 codebook prevously constructed by usng the K-means algorthm [9]. Thus, the output of the feature extracton method conssts of two sequences of dscrete observatons for each dgt: column-based and row-based sequences. 3. Hdden Markov models In the proposed classfer each character class s represented by two HMMs: one based on columns ( λc, λc,..., λ c ) and other based on rows ( λ r, λ r,..., λ r ) of the character mage. These column- and row-based models provde a way of combnng foreground and background features n the zonng scheme as shown n Fgure 4. s s 22222XXXXXXX XXXXXXXXXX XXXXXXXXXXXXX XXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXX XXXXXDDDDDDDDXXXXXXX XXXDDDDDDDDDDDXXXXXX XXXXXX XXXXXX XXXXXXX XXXXXXX XXXXXXX XXXXXXXX XXXXXXXXX XXXXXXXXXX XXXXXXXXXXXXX XXXXXXXXXXXXXX XXXXXXXXXXXXXXX XXXXXXXXDDDXXXXX XXXXXXDDDDDDXXXX XXXXDDDDDDDDXXXX XXXX XXXXX XXXXX XXXXXX XXXXXX XXXXXX XXXXXXX XXXXXXXXXXXX4 333XXXXXXXXXXXXXXXXXXXXXXX4 333XXXXXXXXXXXXXXXXXXXXXX44 333XXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXX XXXXXX XX Fgure 4. Implct zonng scheme provded by combnng column and row-based models The topology of the character models s defned takng nto account the recognton of handwrtten text. Ths means a left-rght model wth number of states defned as descrbed n [10], whch defnes the possble number of states (N) of the HMMs takng nto account duratonal statstcs calculated from the tranng database. Table 1 presents the range (mnmum and maxmum number of states) for each numeral model calculated on the tranng set (50,000 solated dgts 5,000 per class). In addton, the mean length value s also calculated. The fnal number of states of each dgt model was expermentally defned as that correspondng to the mnmum value n Table 1.

4 Table 1. Mnmum, maxmum and mean number of states by dgt class Numeral model Column based models Row based Models Mn Mean Max Mn Mean Max Expermental results Three set of experments were carred out takng nto account solated dgts, solated characters and numeral strngs of dfferent lengths. In all the experments a zerolevel rejecton was used. 4.1 Experments on solated dgts The experments undertaken durng the course of development of the proposed method were done usng solated numerals from the NIST SD19. We use 50,000 numeral samples for tranng, 10,000 for valdaton and 10,000 for testng. A fnal experment was done usng a more robust protocol based on 195,000 samples for tranng, 28,000 for valdaton and 60,089 for testng Evaluaton of the number of states The gap between the number of states usually found n the lterature for HMMs used to represent characters (5 or 6 states) [11,12] and those n Table 1 estmated usng the scheme proposed [10] s very large. For ths reason, we decde at ths tme to evaluate, for the column-based models, confguratons wth 6, 8 and 12 states. Table 2 - Experments consderng dfferent number of states n the numeral HMMs Column models Row models Number of states Vald. (%) Test (%) Vald. (%) Test (%) Mnmum values Mean values Table 2 shows the recognton results consderng dfferent number of states for the column and row numeral models. The best results were obtaned by usng the mnmum values presented n Table 1. The maxmum values were not evaluated snce we have observed a loss n terms of recognton rates for the mean values Evaluaton of the codebook szes The codebook sze was expermentally optmzed. We have evaluated codebooks composed of 64, 128, 192, 256 and 320 entres. The codebook composed of 256 entres provded the best results (see Table 3). The recognton rate of the row models consderng a codebook wth 64 entres were not calculated, snce we have observed that 256 entres provded better results (based on the columnbased models). Table 3 Experments usng dfferent codebook szes Column models Row models Sze Vald. (%) Test. (%) Vald. (%) Test. (%) Combnaton of column and row models The experments have shown that combnng columnand row-based models to represent each dgt class provdes an nterestng recognton performance. In Table 4, experment (a), only the column-based model usng the foreground feature (FF) vector was evaluated. In the experment (b), we observed a sgnfcant mprovement n the recognton performance when we combne foreground and background features n the column-based model. Table 4. Combnaton of column and row models Vald.(%) Testng (%) (a) Column (FF vector) 96,79 94,00 (b) Column (FF + BF vectors) (c) Row (FF + BF vectors) (d) Combnaton (column and row models used n (b) and (c), respectvely) Smlar results were obtaned n the experment (c) for the row-based models. Fnally, the experment (d) has shown that column and row-based models are really complementary. The models were combned by summng the log of ther fnal probablty calculated usng Vterb s algorthm. Table 5 shows the confuson matrx related to the experment (a) presented n Table 4. We can observe many confusons between dgt classes: 0-6, 2-7, 3-5, 4-9, 6-0, 8-6, 9-0 and 9-4.

5 Table 6 shows the fnal confuson matrx related to the experment (d) presented n Table 4. We can observe that the use of complementary nformaton (foreground and background features + column and row models) has shown to be a promsng way to reduce those confusons shown n Table 5. However, there s stll some confuson between classes 2-7, 4-9, and 9-4. Table 5. Confuson matrx: column model (FF vector) Table 6. Confuson matrx: combnaton of column and row models (FF + BF vectors) In a fnal experment consderng solated dgts, we have used a more robust expermental protocol, n whch the database s composed of 195,000 samples for tranng, 28,000 for valdaton and 60,000 for testng. The recognton rate for the testng set was 97.9%. 4.2 Experments on solated characters The solated characters avalable on the NIST SD19 were used n these experments. The hsf_0, hsf_1, hsf_2 and hsf_3 seres were used for tranng. The hsf_7 and hsf_4 seres were used as valdaton and testng sets, respectvely. Table 7 shows the expermental protocol and recognton rates of the proposed method and related works evaluated on the same database. As we can see, t s dffcult to compare snce most of the tme the authors dd not consder 52 classes. We have used the same expermental protocol than Koerch [13] snce the author has consdered the complete expermental protocol. When both, upper and lowercase, are consdered n the same experment the proposed features have shown to be better. The reason s that, the features based on columns and rows have shown to be more sutable to dstngush classes where the dfference s just the sze, such as C and c. Table 7. Expermental protocol and recognton rates on solated characters Class # Tran. # Vald. # Test. Recog. Rates % Proposed method uppercase (26 classes) 37,440 12,092 11, lowercase (26 classes) 37,440 11,578 12, upper/lower (52 classes) 74,880 23,670 23, Koerch [13] uppercase (26 classes) 37,440 12,092 11, lowercase (26 classes) 37,440 11,578 12, upper/lower (52 classes) 74,880 23,670 23, Oh et al [14] 26,000 11, uppercase (26 classes) Dong et al [15] lowercase (26 classes) 23,937 10, Experments on handwrtten numeral strngs We have also used the proposed models and features n a method for numeral strng recognton. It can be categorzed as a segmentaton-free approach that avods a pror segmentaton of the strng nto dgts by usng an mplct segmentaton strategy. In ths method the challenge conssts of fndng the best compromse between segmentaton and recognton. To deal wth ths problem, we propose a two-stage system n [8]. It allows the use of two sets of features and numeral models: one takng nto account of both segmentaton and recognton aspects, and another consderng just the recognton aspects. The frst stage, called the Strng Contextual- Based Stage (SCB), fnds the N best segmentatonrecognton paths for a gven numeral strng. For ths purpose, a dynamc programmng s used to match dgt Hdden Markov Models (HMMs) aganst to a gven strng. The 10 column-based HMMs used n ths stage are traned on solated dgts, but consderng contextual nformaton regardng strng slant and ntra-strng sze varaton. In addton, features extracted from the foreground pxels of the strng mage columns contemplate both segmentaton and recognton processes. The second stage, called Verfcaton Stage, re-ranks the N best segmentatonrecognton paths provded by the frst system stage usng a powerful solated dgt recognzer. The proposed set of features combnng foreground and background nformaton s used n order to mprove the recognton performance of the numeral HMMs. Moreover, 10 addtonal numeral HMMs based on the rows of the numeral mages are combned wth the column-based models durng the dgt recognton process. As we can see, the proposed strng recognton method s totally based on complementary nformaton

6 (foreground + background features, column + row-based models). Ths strategy has been shown to make a sgnfcant contrbuton to strng recognton performance. The recognton rate on 12,802 unknown length strngs composed of 2, 3, 4, 5, 6, and 10 dgts (NIST SD19) were: 94.8%, 91.6%, 91.2%, 88.3%, 89.0%, and 86.9% respectvely. In addton, the recognton rate on 2,069 touchng dgt pars also extracted from NIST SD19 was 89.6%. 5. Concluson In ths paper we have combned complementary features extracted from both foreground and background of character mages. These features were combned n an HMM-based classfer composed of column-based and row-based models. A zonng scheme based on column and row models provdes a way of dvdng the character nto zones wthout makng the features sze nvarant. The experments have shown that HMMs can provde hgh recognton performance close to those provded by the use of neural networks. Ths s very mportant snce HMMs have shown to be more approprate to model handwrtng knowledge related to the nteracton between adjacent characters n words or numeral strngs. Durng the experments t s possble to observe that HMMs and the combnaton of complementary features are a promsng strategy to recognze solated dgts or even to contemplate both segmentaton and recognton aspects n a numeral strng recognton method. Further work can be done n order to develop new features based on structural nformaton. In addton, the system wll be evaluated on cursve words. 6. References [1] Trer O.D., Jan A.K. and Taxt T. Feature Extracton Methods for Character recognton - a Survey. Pattern Recognton, Vol 29, No. 4, pp , [2] Kmura F. and Shrdhar M. Segmentaton-recognton algorthm for handwrtten numeral strngs. Machne Vson Applcatons, No. 5, pp , [3] Cheung K. and Yeung D. A Bayesan Framework for Deformable Pattern Recognton wth Applcaton to Handwrtten Character Recognton. IEEE Trans. on Pattern Analyss and Machne Intellgence, Vol. 20, No. 12, pp , Recognton. Proc. of the Fourteenth Int. Conf. on Pattern Recognton, Vol. I, pp , [6] Heute L., Paquet T., Moreau J.V., Lecourter, Y. and Olver, C. A structural/statstcal feature-based vector for handwrtten character recognton. Pattern Recognton Letters, 19, pp , [7] Olvera L. S., Sabourn R., Bortolozz F., and Suen C.Y. Feature Selecton Usng Mult-Objectve Genetc Algorthms for Handwrtten Dgt Recognton, 16th Int. Conf. on Pattern Recognton, vol I, pp , [8] Brtto Jr, A.S., Sabourn R., Bortolozz F. and Suen C.Y. The Recognton of handwrtten numeral strngs usng a twostage HMM-based method. Int. Journal on Document Analyss and Recognton. Vol. 5, Number 2-3, pp , [9] Makhoul J., Roucos S., and Gsh H. Vector Quantzaton n Speech Codng. Proc. of the IEEE, vol. 73, pp , [10] Wang X. Duratonally constraned tranng of HMM wthout explct state duraton PDF. Inst. of Phonetc Scences, Unversty of Amsterdam, Proceedngs 18, pp , [11] Procter S. and Elms A. J. The recognton of handwrtten dgt strngs of unknown length usng hdden Markov models. Proc. of the Fourteenth Int. Conf. on Pattern Recognton, pp , [12] Rabner, L. R. A Tutoral on Hdden Markov Models and Selected Applcatons n Speech Recognton. Proc. of the IEEE, Vol. 77, No. 2, pp , [13] Koerch, A.L. Large Vocabulary Off Lne Handwrtten Word Recognton. (PhD. Thess) École the Technologa Supéreure (Unversté du Québec). Montreal, 314p, [14] Oh I.S. and Suen C.Y. Dstance features for neural network based recognton of handwrtten characters. Internatonal Journal on Document Analyss and Recognton, 1(2):73 88, [15] Dong J., Krzyzak A., and Suen C.Y. Local learnng framework for recognton of lowercase handwrtten characters. In Proc. Int. Workshop on Machne Learnng and Data Mnng n Pattern Recognton, Lepzg, Germany, pp , [4] Hrano T., Okada Y. and Yoda, F. Structural Character Recognton Usng Smulated Annealng. Proc. of the Fourth Internatonal Conference on Document Analyss and Recognton, Vol. 2, pp , [5] Ca J. and Lu Z. Integraton of Structural and Statstcal Informaton for Unconstraned Handwrtten Numeral

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

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

HYBRID OFF-LINE OCR FOR ISOLATED HANDWRITTEN GREEK CHARACTERS

HYBRID OFF-LINE OCR FOR ISOLATED HANDWRITTEN GREEK CHARACTERS 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

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

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

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

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

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

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

S1 Note. Basis functions.

S1 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 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

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

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

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

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

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

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

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

User Authentication Based On Behavioral Mouse Dynamics Biometrics

User 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 information

Proper Choice of Data Used for the Estimation of Datum Transformation Parameters

Proper 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 information

Smoothing Spline ANOVA for variable screening

Smoothing Spline ANOVA for variable screening Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory

More information

X- Chart Using ANOM Approach

X- 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 information

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task

Term 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 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

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

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

CS 534: Computer Vision Model Fitting

CS 534: Computer Vision Model Fitting CS 534: Computer Vson Model Fttng Sprng 004 Ahmed Elgammal Dept of Computer Scence CS 534 Model Fttng - 1 Outlnes Model fttng s mportant Least-squares fttng Maxmum lkelhood estmaton MAP estmaton Robust

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

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

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

Modeling Inter-cluster and Intra-cluster Discrimination Among Triphones

Modeling Inter-cluster and Intra-cluster Discrimination Among Triphones Modelng Inter-cluster and Intra-cluster Dscrmnaton Among Trphones Tom Ko, Bran Mak and Dongpeng Chen Department of Computer Scence and Engneerng The Hong Kong Unversty of Scence and Technology Clear Water

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

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity Journal of Sgnal and Informaton Processng, 013, 4, 114-119 do:10.436/jsp.013.43b00 Publshed Onlne August 013 (http://www.scrp.org/journal/jsp) Corner-Based Image Algnment usng Pyramd Structure wth Gradent

More information

INTELLECT SENSING OF NEURAL NETWORK THAT TRAINED TO CLASSIFY COMPLEX SIGNALS. Reznik A. Galinskaya A.

INTELLECT SENSING OF NEURAL NETWORK THAT TRAINED TO CLASSIFY COMPLEX SIGNALS. Reznik A. Galinskaya A. Internatonal Journal "Informaton heores & Applcatons" Vol.10 173 INELLEC SENSING OF NEURAL NEWORK HA RAINED O CLASSIFY COMPLEX SIGNALS Reznk A. Galnskaya A. Abstract: An expermental comparson of nformaton

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

Feature Selection for Target Detection in SAR Images

Feature Selection for Target Detection in SAR Images Feature Selecton for Detecton n SAR Images Br Bhanu, Yngqang Ln and Shqn Wang Center for Research n Intellgent Systems Unversty of Calforna, Rversde, CA 95, USA Abstract A genetc algorthm (GA) approach

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

SURFACE PROFILE EVALUATION BY FRACTAL DIMENSION AND STATISTIC TOOLS USING MATLAB

SURFACE PROFILE EVALUATION BY FRACTAL DIMENSION AND STATISTIC TOOLS USING MATLAB SURFACE PROFILE EVALUATION BY FRACTAL DIMENSION AND STATISTIC TOOLS USING MATLAB V. Hotař, A. Hotař Techncal Unversty of Lberec, Department of Glass Producng Machnes and Robotcs, Department of Materal

More information

Feature-based image registration using the shape context

Feature-based image registration using the shape context Feature-based mage regstraton usng the shape context LEI HUANG *, ZHEN LI Center for Earth Observaton and Dgtal Earth, Chnese Academy of Scences, Bejng, 100012, Chna Graduate Unversty of Chnese Academy

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

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

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

Detection of hand grasping an object from complex background based on machine learning co-occurrence of local image feature

Detection of hand grasping an object from complex background based on machine learning co-occurrence of local image feature Detecton of hand graspng an object from complex background based on machne learnng co-occurrence of local mage feature Shnya Moroka, Yasuhro Hramoto, Nobutaka Shmada, Tadash Matsuo, Yoshak Shra Rtsumekan

More information

D. Barbuzzi* and G. Pirlo

D. Barbuzzi* and G. Pirlo Int. J. Sgnal and Imagng Systems Engneerng, Vol. 7, No. 4, 2014 245 bout retranng rule n mult-expert ntellgent system for sem-supervsed learnng usng SVM classfers D. Barbuzz* and G. Prlo Department 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

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

Comparison Study of Textural Descriptors for Training Neural Network Classifiers

Comparison 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 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

Classifying Acoustic Transient Signals Using Artificial Intelligence

Classifying 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 information

Sixth Indian Conference on Computer Vision, Graphics & Image Processing

Sixth Indian Conference on Computer Vision, Graphics & Image Processing Sxth Indan Conference on Computer Vson, Graphcs & Image Processng Incorporatng Cohort Informaton for Relable Palmprnt Authentcaton Ajay Kumar Bometrcs Research Laboratory, Department of Electrcal Engneerng

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation 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 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

Collaboratively Regularized Nearest Points for Set Based Recognition

Collaboratively Regularized Nearest Points for Set Based Recognition Academc Center for Computng and Meda Studes, Kyoto Unversty Collaboratvely Regularzed Nearest Ponts for Set Based Recognton Yang Wu, Mchhko Mnoh, Masayuk Mukunok Kyoto Unversty 9/1/013 BMVC 013 @ Brstol,

More information

Mining Image Features in an Automatic Two- Dimensional Shape Recognition System

Mining Image Features in an Automatic Two- Dimensional Shape Recognition System Internatonal Journal of Appled Mathematcs and Computer Scences Volume 2 Number 1 Mnng Image Features n an Automatc Two- Dmensonal Shape Recognton System R. A. Salam, M.A. Rodrgues Abstract The number of

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

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

Optimizing Document Scoring for Query Retrieval

Optimizing Document Scoring for Query Retrieval Optmzng Document Scorng for Query Retreval Brent Ellwen baellwe@cs.stanford.edu Abstract The goal of ths project was to automate the process of tunng a document query engne. Specfcally, I used machne learnng

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

PERFORMANCE EVALUATION FOR SCENE MATCHING ALGORITHMS BY SVM

PERFORMANCE EVALUATION FOR SCENE MATCHING ALGORITHMS BY SVM PERFORMACE EVALUAIO FOR SCEE MACHIG ALGORIHMS BY SVM Zhaohu Yang a, b, *, Yngyng Chen a, Shaomng Zhang a a he Research Center of Remote Sensng and Geomatc, ongj Unversty, Shangha 200092, Chna - yzhac@63.com

More information

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr)

Helsinki 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 information

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

A 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 information

TN348: Openlab Module - Colocalization

TN348: 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 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

Analysis of Continuous Beams in General

Analysis of Continuous Beams in General Analyss of Contnuous Beams n General Contnuous beams consdered here are prsmatc, rgdly connected to each beam segment and supported at varous ponts along the beam. onts are selected at ponts of support,

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

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

Arabic Handwritten Word Recognition based on Dynamic Bayesian Network

Arabic Handwritten Word Recognition based on Dynamic Bayesian Network 1024 The Internatonal Arab Journal of Informaton Technology Vol. 13, o. 6B, 2016 Arabc Handwrtten Word Recognton based on Dynamc Bayesan etwork Khaoula Jayech, Mohamed Al Mahjoub, and ajoua Ben Amara Department

More information

Three supervised learning methods on pen digits character recognition dataset

Three supervised learning methods on pen digits character recognition dataset Three supervsed learnng methods on pen dgts character recognton dataset Chrs Flezach Department of Computer Scence and Engneerng Unversty of Calforna, San Dego San Dego, CA 92093 cflezac@cs.ucsd.edu Satoru

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

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

Parallel Numerics. 1 Preconditioning & Iterative Solvers (From 2016)

Parallel Numerics. 1 Preconditioning & Iterative Solvers (From 2016) Technsche Unverstät München WSe 6/7 Insttut für Informatk Prof. Dr. Thomas Huckle Dpl.-Math. Benjamn Uekermann Parallel Numercs Exercse : Prevous Exam Questons Precondtonng & Iteratve Solvers (From 6)

More information

Intelligent Information Acquisition for Improved Clustering

Intelligent Information Acquisition for Improved Clustering Intellgent Informaton Acquston for Improved Clusterng Duy Vu Unversty of Texas at Austn duyvu@cs.utexas.edu Mkhal Blenko Mcrosoft Research mblenko@mcrosoft.com Prem Melvlle IBM T.J. Watson Research Center

More information

Title: A Novel Protocol for Accuracy Assessment in Classification of Very High Resolution Images

Title: A Novel Protocol for Accuracy Assessment in Classification of Very High Resolution Images 2009 IEEE. Personal use of ths materal s permtted. Permsson from IEEE must be obtaned for all other uses, n any current or future meda, ncludng reprntng/republshng ths materal for advertsng or promotonal

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

Biostatistics 615/815

Biostatistics 615/815 The E-M Algorthm Bostatstcs 615/815 Lecture 17 Last Lecture: The Smplex Method General method for optmzaton Makes few assumptons about functon Crawls towards mnmum Some recommendatons Multple startng ponts

More information

Slant normalization of handwritten numeral strings

Slant normalization of handwritten numeral strings Slant normalization of handwritten numeral strings Alceu de S. Britto Jr 1,4, Robert Sabourin 2, Edouard Lethelier 1, Flávio Bortolozzi 1, Ching Y. Suen 3 adesouza, sabourin@livia.etsmtl.ca suen@cenparmi.concordia.ca

More information

Multi-View Face Alignment Using 3D Shape Model for View Estimation

Multi-View Face Alignment Using 3D Shape Model for View Estimation Mult-Vew Face Algnment Usng 3D Shape Model for Vew Estmaton Yanchao Su 1, Hazhou A 1, Shhong Lao 1 Computer Scence and Technology Department, Tsnghua Unversty Core Technology Center, Omron Corporaton ahz@mal.tsnghua.edu.cn

More information

A NEW FUZZY C-MEANS BASED SEGMENTATION STRATEGY. APPLICATIONS TO LIP REGION IDENTIFICATION

A NEW FUZZY C-MEANS BASED SEGMENTATION STRATEGY. APPLICATIONS TO LIP REGION IDENTIFICATION A NEW FUZZY C-MEANS BASED SEGMENTATION STRATEGY. APPLICATIONS TO LIP REGION IDENTIFICATION Mhaela Gordan *, Constantne Kotropoulos **, Apostolos Georgaks **, Ioanns Ptas ** * Bass of Electroncs Department,

More information

RECOGNIZING GENDER THROUGH FACIAL IMAGE USING SUPPORT VECTOR MACHINE

RECOGNIZING GENDER THROUGH FACIAL IMAGE USING SUPPORT VECTOR MACHINE Journal of Theoretcal and Appled Informaton Technology 30 th June 06. Vol.88. No.3 005-06 JATIT & LLS. All rghts reserved. ISSN: 99-8645 www.jatt.org E-ISSN: 87-395 RECOGNIZING GENDER THROUGH FACIAL IMAGE

More information

A Statistical Model Selection Strategy Applied to Neural Networks

A 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 information

Active Contours/Snakes

Active Contours/Snakes Actve Contours/Snakes Erkut Erdem Acknowledgement: The sldes are adapted from the sldes prepared by K. Grauman of Unversty of Texas at Austn Fttng: Edges vs. boundares Edges useful sgnal to ndcate occludng

More information

Fast Feature Value Searching for Face Detection

Fast Feature Value Searching for Face Detection Vol., No. 2 Computer and Informaton Scence Fast Feature Value Searchng for Face Detecton Yunyang Yan Department of Computer Engneerng Huayn Insttute of Technology Hua an 22300, Chna E-mal: areyyyke@63.com

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

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

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem 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 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

Online Handwritten Cursive Word Recognition Using Segmentation-free and Segmentation-based Methods

Online Handwritten Cursive Word Recognition Using Segmentation-free and Segmentation-based Methods 3th Asan Conference on Pattern Recognton (ACPR2015), November 3-6, 2015, Kuala Lumpur, Malasa. Onlne Handwrtten Cursve Word Recognton Usng Segmentaton-free and Segmentaton-based Methods Blan Zhua, Art

More information

Multiple Frame Motion Inference Using Belief Propagation

Multiple Frame Motion Inference Using Belief Propagation Multple Frame Moton Inference Usng Belef Propagaton Jang Gao Janbo Sh The Robotcs Insttute Department of Computer and Informaton Scence Carnege Mellon Unversty Unversty of Pennsylvana Pttsburgh, PA 53

More information

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009. Farrukh Jabeen Algorthms 51 Assgnment #2 Due Date: June 15, 29. Assgnment # 2 Chapter 3 Dscrete Fourer Transforms Implement the FFT for the DFT. Descrbed n sectons 3.1 and 3.2. Delverables: 1. Concse descrpton

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

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

Audio Content Classification Method Research Based on Two-step Strategy

Audio Content Classification Method Research Based on Two-step Strategy (IJACSA) Internatonal Journal of Advanced Computer Scence and Applcatons, Audo Content Classfcaton Method Research Based on Two-step Strategy Sume Lang Department of Computer Scence and Technology Chongqng

More information

MOTION BLUR ESTIMATION AT CORNERS

MOTION BLUR ESTIMATION AT CORNERS Gacomo Boracch and Vncenzo Caglot Dpartmento d Elettronca e Informazone, Poltecnco d Mlano, Va Ponzo, 34/5-20133 MILANO boracch@elet.polm.t, caglot@elet.polm.t Keywords: Abstract: Pont Spread Functon Parameter

More information

A Multilevel Algorithm for the Network Design Problem

A Multilevel Algorithm for the Network Design Problem A Multlevel Algorthm for the Network Desgn Problem Hdeson A. Slva 1,2, Alceu S. Brtto Jr 2, Luz E. S. Olvera 3, Alessandro L. Koerch 2,3 1 Companha Paranaense de Energa(Copel), 2 Pontfíca Unversdade Católca

More information

Capturing Global and Local Dynamics for Human Action Recognition

Capturing Global and Local Dynamics for Human Action Recognition 2014 22nd Internatonal Conference on Pattern Recognton Capturng Global and Local Dynamcs for Human Acton Recognton Sq Ne Department of Electrcal, Computer and System Engneerng Rensselaer Polytechnc Insttute

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

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

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

Wishing you all a Total Quality New Year!

Wishing 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 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