Visual Thesaurus for Color Image Retrieval using Self-Organizing Maps

Size: px
Start display at page:

Download "Visual Thesaurus for Color Image Retrieval using Self-Organizing Maps"

Transcription

1 Vsual Thesaurus for Color Image Retreval usng Self-Organzng Maps Chrstopher C. Yang and Mlo K. Yp Department of System Engneerng and Engneerng Management The Chnese Unversty of Hong Kong, Hong Kong ABSTRACT The technque of searchng n content-based mage retreval has been actvely studed n recent years. However, ths technque cannot gve user an overvew of the database. In ths paper, we propose a browsng technque usng Kohonen s Self-Organzng Map to retreve general color mage database effectvely. Both chromatc and textural feature of mages are analyzed to represent the content of mages. Keywords: Vsual database, Content-based Retreval, Image Browsng, Self-Organzng Map, Color Hstogram, Gabor Flter 1. INTRODUCTION Recently, many approaches for content-based mage retreval (CBIR) have been developed; for example, IBM s Query by Image Content (QBIC), and UC Santa Barbara s Alexandra Dgtal Lbrary (ADL) proect. Most of them use the technque of query-by-example, a searchng approach to fnd mages that are smlar to the gven example mage. Smlarty between mages s quantfed n terms of some global or local features, such as colour, texture, shape, pattern, or specfc spatal area of the above. The feature selecton and matchng technques are vtal to searchng thus much research worked on these technques. However, there are some serous lmtatons on usng searchng solely. For example, query-by-example technques often generate results of a relatvely small number of mages, whch may not be nterested by the user; as a result, the user may not be able to make further queryng. In contrast, browsng envronment offer an alternate approach to the problem, but t has receved relatvely lttle attenton. Accordng to Craverl et al. [1], browsng s a technque, or a process, that the users vew nformaton rapdly and decde whether the content s relevant to ther needs. The user can get an overvew or summary of content quckly and then focus on partcular sectons of nterest. Craver et al. modeled a general process of mage browsng. The frst step s to extract the relatonshp among the mages. The second step s to fnd representatve nstances of mages durng browsng. The fnal step s to vsualze the mages wth ntutve presentaton of the relatonshp among mages. Recent research n mages browsng uses dfferent data structure to organze mages n a meanngful way to present the relatonshp among mages. Craver et al. descrbed a technque based on multple space-fllng curves. Chen et al. [] ntated a new approach called actve browsng. The research made use of ther prevous work on smlarty pyramds. Those technques provde an attractve feature that the process of nserton of new mages s computatonally fast; complete re-ndexng s not necessary. We nvestgate the unsupervsed artfcal network, Kohonen s self- organzng map (SOM) [3, 4], for mage browsng. SOM reduces hgh-dmensonal nput sgnal space nto low-dmensonal space whle the map preserves the relatonshp among the nput sgnals. Zhang et al. [5] addressed the mportance of developng effectve ndexng scheme for mage database searchng. They studed an ndexng scheme that apples SOM approach. Three texture features MRSAR, coarseness and gray hstogram were used for ndexng monochrome textural mages and color hstogram was used for ndexng general color mages. They developed a set of herarchcal self-organzng maps (HSOMs) to construct an ndex tree, whch provdes a space for searchng. The evaluaton of performance s calculated by retreval rate. Han et al. [6] uses SOM for mage retreval. Ther system used SOM for generatng maps for human browsng but not for machne ndexng. However, ther system s restrcted to handle mages of obects, from whch shape features such as roundness, rectangularty, elptcty, eccentrcty and bendng energy are extracted, where only the luance nformaton s consdered. Besdes, they used a set of randomly selected mages to label the regons n the map, whch s not necessary representng the most sgnfcant features of the node. In ths paper, we present an approach to provde a mage-browsng nterface for human to retreve general colour mage database usng SOM. We have tested advanced chromatc and texture features to generate maps and constructed three methods to label the resultant maps.. ALGORITHM Our proposed algorthm can be dvded nto two man sectons: mage representaton and self-organzng map generaton. The former secton transforms the mage database nto representaton, whch can be used for evaluatng smlarty among the mages. The latter secton generates SOM usng the above nformaton. Image Representaton There are a number of colour coordnate systems for mages. Tradtonally colour mages are represented n RGB format when processed n computer systems. However, ths colour coordnate system does not match wth human percepton of colour. Colour percepton research generally beleve that human recognze colour by hue and saturaton. Snce ths research s amed at provdng an nterface for human, we have adopted LHS color coordnate system that corresponds to human color percepton. The process of the proposed algorthm for mage representaton s shown as follows.

2 Raw Image Fgure 1. The process for mage representaton Feature Extracton: In order to extract the content of mages n the database, we need to convert raw data of mages nto some presentable and comparable features. There are varous knds of features of mages, for example, colour, texture, shape and pattern. In ths thess, we have chosen colour and texture features because the type of database s general colour mages, shape and pattern features are not sutable. For the colour feature, we used the hue and saturaton component n LHS color coordnate system. Although all three components are necessary to represent a colour, luance of LHS space s omtted n our colour feature. It s because mages captured n the same scene at dfferent angles wll affect ts hue and saturaton dstrbuton only slghtly but luance dstrbuton wll be easly affected due to lghtng of the envronment. Besdes, the texture feature s analyzed by the luance component only. Therefore, the process s denoted as: For each mage, Feature Extracton RGB Extracton LHS Transformaton Feature Analyss Chromatc Analyss Textural Analyss Image Representaton 1. Extract RGB values of each pxel n the mage. Convert the RGB values nto LHS color coordnate system 3. Perform textural analyss wth the L value 4. Perform chromatc analyss wth the H and S values 5. Combne both analyss result to represent the mage feature RGB L HS Textural Analyss Chromatc Analyss Image Feature The hue component vares from 0 to 360 degrees and the saturaton component vares from 0 to 1. In order to buld the hstogram, the hue and saturaton values are quantzed nto several levels. We have chosen 10 levels for each component, so that a 10 by 10 two-dmensonal hstorgram h[10,10] for each mage s bult n our system. A Pxel wth 0 h < 36 and 0 s < 1 s sorted n the frst bn h[1,1] and so on. Textural Analyss: There s a long hstory for texture analyss research. Accordng to Manunath et al. [8], texture analyss algorthms uses varous technques, ncludng random feld model and multresoluton flterng technques such as wavelet transform. In the research of Manunath et al., the experment results of a number of texture features are compared. The texture features ncludes the conventonal pyramd-structured wavelet transform (PWT) features, tree-structured wavelet transform (TWT) features, and the multresoluton smultaneous autoregressve model (MR-SAR) features and the Gabor wavelet features. The expermental results show that Gabor features gve the best performance. In another recent research carred out by Chen et al. [9], they compared four flterng technques ncludng Fourer transform, spatal flter, Gabor flter and wavelet transform for texture dscraton. The expermental results also show that Gabor flter generally gves the best performance. However, Chen et al. stated that the executon tme of Gabor flter s the longest out of the four features. Scne texture analyss are generated once only n ths knd of mage retreval applcaton, the dsadvantage of longer executon tme of generatng Gabor features s not crtcal. Therefore, we adopt Gabor flter for textural analyss n ths research due to ts best performance. In the followng, we wll revew the Manunath et al. s method of usng Gabor flter for texture feature extracton. A two-dmensonal Gabor functon can be wrtten as: Fgure. The nformaton flow of the process for mage representaton x y g ( 1 1 x, y) = exp + + πwx πσ xσ y σ x σ y () Chromatc Analyss: We adopt the concept of colour ndexng method proposed by Swan et al. [7]. Ths method uses hstogram of ntensty values to represent the colour dstrbuton. Ths knd of hstogram captures the global chromatc nformaton of an mage. The advantage of ths method s that the hstogram s nvarant under translaton and rotaton about the vew axs. Despte changes n vew, change n scale, and occluson, the hstogram only changes slghtly. As mentoned before, we use H and S values for chromatc analyss. The algorthm s to buld a two-dmensonal hstogram wth one axs for hue and another for saturaton. The defnton of HS hstogram s: Hstogram h, s [ h, s] = N Pr( H = h, S = s) where H and S s the hue and saturaton channels, N s the number of pxels n the mage. (1) Then a self-smlar flter dctonary can be obtaned as a mother Gabor Wavelet G(x, y) by approprate dlatons and rotatons of Eq. () as: G = a where S m G θ, θ ) ( x y h = heght of hsde = (h -1) mage, w /, wsde θ = ( x hsde)cos( + ( y wsde)sn( x θ = ( x hsde)sn( + ( y wsde)cos( y a > 1, m,n are ntegers = wdth of mage, = (w - Gven an mage wth luance ( x y) 1) / I,, a Gabor decomposton can be obtaned by multplyng the luance by the magntude of the Gabor Wavelet: (3)

3 W (4) The mean and standard devaton of the magntude of the transform coeffcent are used to represent the texture feature for classfcaton and retreval purpose: The Gabor feature vector s constructed by usng σ as feature components: f ( x, y) = I( x, y) G G r + ( x y) W, dxdy µ = h w σ [ µ σ µ σ µ σ ] = ( S 1)( K 1) ( S 1)( K 1) ( W ( x, y) ( x, y ) dxdy = µ (5) (6) µ and where S s number of scales and K s number of orentaton. In the followng experment, we use S=3 and K=4. Self-Organzng Map After generatng mage representaton, we need to vsualze the representaton of all mages n database by categorzaton. Categorzaton s a process whch nvolves groupng tems of smlar nature. There are several clusterng algorthms avalable, such as K-means algorthm, sngle-lnk clusterng and complete-lnk clusterng (based on mum-spannng tree). A newer connectonst approach called Self-Organzng Map (SOM) s developed by Kohonen n SOM s supposed to reduce the hgh dmensonal nput space nto lower dmensonal space, usually two-dmensonal. The resultant feature mappng s a non-lnear proecton from the nput space. The topologcally close nodes n the map are senstve to the nputs that are smlar. Structure: Accordng to Kohonen, the structure of SOM s typcally an array of nodes (neurons) arranged n two-dmensonal space. The nodes can be arranged n varous ways, for example, rectangular, hexagonal, or even rregular. Rectangular arrangement s used n ths research, as t s most sutable for vsualzaton of a set of mages. The structure of SOM used n ths research s llustrated as follows. Input Pattern Input Layer Kohonen Layer Fgure 3. The structure of D SOM (7) The nput patterns are the feature vectors of mages n the database. Each node n nput layer (component of feature vector) s fully connected to the Kohonen Layer. Tranng: The notatons and tranng algorthm of SOM used n ths research s as follows. I Set of all mages n database x Feature vector of mage I M Set of all nodes n Kohonen Layer w (t) Weght vector of node M at tme t R Set of mages that are mapped to node M v Label mage of node M 1. Randomze w ( 0), M. For each teraton,.1 Shuffle the presentaton order of mages. For each mage n the orderng lst,..1 Fnd a wnnng node c.. Update the wnnng node c and ts neghbors.3 Contnue tranng wth decreased learnng rate and reduced area of neghborhood functon 3. For each I, put nto R Frstly, the weghts of all nodes are assgned wth a vector of random values. If there are dentcal weghts of node, there may be a case that there are two wnnng nodes. To prevent ths stuaton, the random process s used. In each teraton, all mages wll be presented to the system. However, f the sequence of presentaton of nodes s fxed, the map wll be nfluence heavly by the frst few mages. Therefore, the order of presentaton s shuffled n the begnnng of each teraton to mze ths effect. When an mage s presented to the system, the feature vector x s compared wth all weght vector w (t) of all nodes n Kohonen Layer. The most smlar node, called the wnnng node, can be found. Smlarty s defned by Eucldean dstance. Denotng the wnnng node as c (t), fndng a wnnng node s formulated as: c ( t) = arg x w M () t Then the wnnng node c (t) and ts neghbors are updated. The obectve of updatng s to make the weghts of the wnnng node and ts neghbors become more smlar to that nput vector. The formula and the llustraton of the vectors are shown as follow. w ( t + 1) = w () t + ηh () t [ x w () t ] c where 0 < η < 1 s the learnng factor and h c () t s the neghborhood functon. For convergence t s necessary η h c () t 0 when t. In ths research, the learnng factor η s a lnearly decreasng functon, of whch the ntal and fnal values are specfed explctly. In ths research, we use a constant weghtng for the neghborhood functon. The area of t dshed wth respect to tme. (8) (9)

4 h c () t 1 = 0 f N c ( t) f N ( t) c (10) where N c (t) refers to a neghborhood set of array ponts around node c at tme t. It has been depcted as: y = k k R R x arg x y f R > 0 v = R undefned otherwse (14) 3. Label by result-smlarty: For each node n the map, fnd an mage R such that the Eucldean dstance between w and x s mal. c N c (t 1 ) N c (t ) N c (t 3 ) v arg x w f R > 0 R undefned otherwse = 3. Illustraton (15) Fgure 4. Neghborhood set (t 1 < t < t 3 ) We defned the length of the sde of square shape neghborhood set as: length (,3) () t = max max( SOM, SOM )( 1 t) wdth heght (11) At the begnnng (t = 0), the neghborhood set covers at most the whole SOM. Its sze dshed wth respect to t untl the length equals to 3. After the network s traned through a number of teratons, a converged SOM s generated. The resultant SOM contans weghts whch represent the dstrbuton of nput space. Fnally, feature vectors of mages are mapped to the nodes by the mum Eucldean dstance, formulated as: k = arg x w R = M { k = } () t where k denotes the node whch mage s mapped to, and R denotes the set of mages mapped to node. (1) Labelng: After mappng the feature vectors, nodes n the SOM may contan zero or many mages. To vsualze the SOM, we need to label each node by a representatve mage. Three algorthms of labelng are tested n ths research as follows. 1. Label by smlarty: For each node n the map, fnd an mage I as label such that the Eucldean dstance between x and w s mal. v = arg x w I (13). Label by result-mean: For each node n the map, calculate the mean y of weght vectors of mages n R. Next, fnd the mage wth mal Eucldean dstance to the mean as label. We mplemented the system n Java (JDK 1.1.8) and run t on a Pentum II-33 PC. There are two modules n the system. The frst one s feature module, whch extract chromatc and textural feature of mages n the database. It also ncludes user-nterface for retrevng feature nformaton of each mage. A smple query-by-example s mplemented n ths module for testng the performance of the features. The second one s SOM module whch trans SOMs usng feature databases. It contans a user-nterface to vsualze the labeled result for browsng the mages. In the followng, we wll evaluate SOMs usng dfferent mage features and labelng method. Feature dfferences We have used a database wth 500 colour mages (640x480x4bts) n the followng experments. Fgure 5 (a-c) shows a 10x10 SOM traned wth dfferent features. Three SOMs are traned for 50 teratons 1. Learnng rate begns at 0.5 and ends at Resultng maps are labeled by result-smlarty. The number n left-lower corner n each node ndcates the number of mages n the node ( R ). After selectng a node n the map by user, the rght colu dsplays the mages (R ) mapped to the selected node. All the resultng maps can provde an overvew of the database. And more mportantly, they contan rch nformaton. They vsualze the underlyng structure of the feature space by dsplayng topologcally regons of maor concepts n the mage space and the dstrbuton over the concept regons. We wll evaluate the three SOMs n the followng. SOM n fgure 5(a) only uses HS-hstogram, a chromatc feature. It vsualzes the dstrbuton of color of mages. For example, mages wth blues are put on the rght-bottom of map. However, t cannot dfferentate mages wth smlar colors but dfferent textures. For example, some wnter scenery photos and some obect photos both are almost black and whte, therefore the concepts of these two categorzes of mages are not clearly dstngushed n the map. Also, obects wth dfferent colors lke tenns-ball and baseball are separate apart. 1 The term teraton n our algorthm s defned as number of presentaton of all mages (see secton.). However, most SOM algorthms defne teraton as number of presentaton of nput patterns. Usng the latter defnton, there are 500 x 50 = 5000 teratons n ths experment

5 Gabor flter, a textural feature, s used to tran the SOM shown n fgure 5(b). Contrary to the former SOM, t can separate obects from scenery photos perfectly but photos wth smlar colors are not grouped together. For nstance, photos wth blues and greens are n dfferent regons n the map. The fnal one combnes HS-hstogram and Gabor flter. Advantages of chromatc and textural features are found n ths map. Obects and scenery photos are separate apart and mages wth smlar colors are close together. Of course the performance s not as good as fgure 5(a) n terms of chromatc dfferentate performance, and fgure 5(b) n terms of textural dfferentate performance. Nevertheless, t provdes a balance of usng both features. Labelng dfferences We have also compared dfferent labelng methods. Fgures 5(c) and 6(a-b) dsplays the same SOM traned by same parameters n secton 3.1 but labeled wth dfferent methods. 5(a) HS-Hstogram The map n fgure 6(a) s labeled by smlarty, that s, usng the most smlar mage n the database to represent the node. It vsualzes best the nformaton of the weghts traned n the map thus present an overvew of dstrbuton of mages n the database and the map s generally very contnuous. However, the label mage of a node may not be an mage mapped to that node. And there are cases of duplcate representatve mages. If t s used as a user-nterface for browsng, users may confuse f ther nterested mage exsts repeatedly throughout the map; they do not know whch node they should further explore. Fgure 6(b) shows a map labeled by result-mean. The label mage s chosen from the most smlar mage to the mean of features vector of mages mapped to that node. It prevents the problem ust mentoned of labelng by smlar and t fnds the best mages to represent mages mapped to that node. However, the label mages are not the most smlar one to the weghts of the nodes thus the map s vsually more fragmented. 5(b) Gabor Flter The last map s shown n fgure 5(c), whch s labeled by result-smlarty. It has some advantages of labelng by smlarty and labelng by result-mean. It provdes a balance between representng weghts of nodes and representng the features of mages mapped to the nodes. Two mportant propertes of result-mean and result- smlarty methods are that the label mage must be ncluded n mages mapped to the node and duplcate label mage wll not exst. 4. CONCLUSION We have proposed an approach for effcent colour mage retreval. Image feature and SOM are used to vsualze a summary of all mages n the database. Experments show that mage features can be combned to balance the advantages of dfferent features. And also, SOM labeled by result-nearest s most sutable for representng both the map and the mages mapped to the nodes. 5(c) Combnaton of HS-hstogram and Gabor Flter Fgure 5(a-c): 10x10 SOMs traned by dfferent features (all three maps are labeled by result-smlarty)

6 exstng SOM to adapt the new mages. Nevertheless t cannot apply for HSOM/M-SOM approach; modfcatons of top-level SOM make lower-level SOM to be re-traned. We wll contnue to study on ths feld n order to solve ths problem. A problem we have encountered s quanttatve evaluaton of the results. We clam that the usefulness of maps can be evaluated by practcal tests only. Therefore, we wll setup human experment for testng the performance of human n dfferent varables. As SOM provde a vsual overvew of the database, once subects have used a SOM for a perod, we clam that they wll have a better performance due to mplct learnng. 6(a) Label by Smlarty Eventually, browsng and searchng are not compettve technques. We wll work on a proposal whch combnes the two maor approaches. More features and new combnatons of features wll be tested. For nstance, text-annotaton s a useful feature for manual categorzaton of mage database. 5. ACKNOWLEDGEMENT Ths proect s supported by the Earmarked Grant for Research of the Research Grants Councl of Hong Kong, CUHK E. 6. REFERENCES 6(b) Label by Result-Mean Fgure 6(a-b): 10x10 SOMs labeled by dfferent methods (both are traned by combned features) The results of experments are encouragng. We thnk that SOM s an deal technque for generatng maps for mage browsng snce t can reflect the relatonshp among all mages n the database. However, there are some restrctons of current approach usng SOM. Frst, f the number of mages s large, the number of mages mapped n one node wll be large. It s neffcent for user to browse. An mmedate derved approach to ths problem s to generate SOMs for the mages mapped to each node. We can apply ths approach recursvely untl the number of mages n the node s less than a specfc amount. Actually, ths approach s smlar to the Herarchcal Self-Organzng Map (HSOM) from Zhang et al. and the Multlayered Self-Organzng Feature Map (M-SOM) from Chen et al. [10]. The purpose of these two research are vsual ndexng and nternet categorzaton respectvely. Ths knd of SOM varatons can be appled to our research also. We wll have to desgn approprate algorthms for labelng. Another restrcton s lack of dynamc mantenance. Inserton or deleton mages needs to re-tran the SOM. For a normal sngle layer SOM, we can apply addtonal tranng teratons to the [1] S.Craver, B.-L. Yeo, and M.M. Yeung, Image browsng usng data structure based on multple space-fllng curves. To appear n the Procceedngs of the Thrty-Sxth Aslomar Conference on Sgnals, Systems, and Computers, pp (Pacfc Grove, CA), November [] J.-Y. Chen, C.A. Bouman and J. C. Dalton, Actve Browsng usng Smlarty Pyramds, Proc. Of IS&T / SPIE Conference on Storage and Retreval for Image and Vdeo Database VII, vol. 3656, pp , (San Jose, Calforna), January [3] T. Kohonen, Self-Organzaton and Assocatve Memory, Sprnger Seres n Informaton Scence, vol. 8, [4] T. Kohonen, Self-Organzng Maps, Sprnger Seres n Informaton Scence, vol. 30, [5] H. Zhang and D. Zhong, A Scheme for Vsual Feature based Image Indexng, n Proc. Of IS&T / SPIE Conference on Storage and Retreval for Image and Vdeo Database III, vol. 40, pp , (San Jose, CA), February [6] K. A. Han, J. C. Lee and C. J. Hwang, Image Clusterng usng Self-organzng feature map wth Refnement, Proceedngs IEEE Internatonal Conference on Neural Networks, Vol. 1, pp , [7] M. Swan and D. Ballard, Color Indexng, Internatonal Journal of Computer Vson, 7:1, pp 11-3, [8] B. S. Manunath and W. Y. Ma, Texture Features for Browsng and Retreval of Image Data, IEEE Transactons on Pattern Analyss and Machne Intellgence, Vol. 8, No. 18, pp , August [9] C. C. Chen and C. C. Chen, Flterng Methods for Texture Dscraton, Pattern Recognton Letters, vol. 0, pp , [10] H. Chen, C. Schuffels and R. Orwg, Internet Categorzaton and Search: A Self-Organzng Approach, Journal of Vsual Communcaton and Image Representaton, Vol. 7, No.1, pp , March 1996.

7

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

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

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

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

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

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

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

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

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

Object-Based Techniques for Image Retrieval

Object-Based Techniques for Image Retrieval 54 Zhang, Gao, & Luo Chapter VII Object-Based Technques for Image Retreval Y. J. Zhang, Tsnghua Unversty, Chna Y. Y. Gao, Tsnghua Unversty, Chna Y. Luo, Tsnghua Unversty, Chna ABSTRACT To overcome the

More information

Lecture 5: Multilayer Perceptrons

Lecture 5: Multilayer Perceptrons Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented

More 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

COMPLEX WAVELET TRANSFORM-BASED COLOR INDEXING FOR CONTENT-BASED IMAGE RETRIEVAL

COMPLEX WAVELET TRANSFORM-BASED COLOR INDEXING FOR CONTENT-BASED IMAGE RETRIEVAL COMPLEX WAVELET TRANSFORM-BASED COLOR INDEXING FOR CONTENT-BASED IMAGE RETRIEVAL Nader Safavan and Shohreh Kasae Department of Computer Engneerng Sharf Unversty of Technology Tehran, Iran skasae@sharf.edu

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

Searching Large Image Databases using Color Information

Searching Large Image Databases using Color Information Searchng Large Image Databases usng Color Informaton Ioan Racu CMSC350: Artfcal Intellgence Wnter Quarter 2004 Department of Computer Scence Unversty of Chcago racu@cs.uchcago.edu ABSTRACT The goal of

More information

1. Introduction. Abstract

1. Introduction. Abstract Image Retreval Usng a Herarchy of Clusters Danela Stan & Ishwar K. Seth Intellgent Informaton Engneerng Laboratory, Department of Computer Scence & Engneerng, Oaland Unversty, Rochester, Mchgan 48309-4478

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

Image Alignment CSC 767

Image Alignment CSC 767 Image Algnment CSC 767 Image algnment Image from http://graphcs.cs.cmu.edu/courses/15-463/2010_fall/ Image algnment: Applcatons Panorama sttchng Image algnment: Applcatons Recognton of object nstances

More information

Semantic Image Retrieval Using Region Based Inverted File

Semantic Image Retrieval Using Region Based Inverted File Semantc Image Retreval Usng Regon Based Inverted Fle Dengsheng Zhang, Md Monrul Islam, Guoun Lu and Jn Hou 2 Gppsland School of Informaton Technology, Monash Unversty Churchll, VIC 3842, Australa E-mal:

More information

A Knowledge Management System for Organizing MEDLINE Database

A Knowledge Management System for Organizing MEDLINE Database A Knowledge Management System for Organzng MEDLINE Database Hyunk Km, Su-Shng Chen Computer and Informaton Scence Engneerng Department, Unversty of Florda, Ganesvlle, Florda 32611, USA Wth the exploson

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

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

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

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

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

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

Background Removal in Image indexing and Retrieval

Background Removal in Image indexing and Retrieval Background Removal n Image ndexng and Retreval Y Lu and Hong Guo Department of Electrcal and Computer Engneerng The Unversty of Mchgan-Dearborn Dearborn Mchgan 4818-1491, U.S.A. Voce: 313-593-508, Fax:

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

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems A Unfed Framework for Semantcs and Feature Based Relevance Feedback n Image Retreval Systems Ye Lu *, Chunhu Hu 2, Xngquan Zhu 3*, HongJang Zhang 2, Qang Yang * School of Computng Scence Smon Fraser Unversty

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

Combination of Color and Local Patterns as a Feature Vector for CBIR

Combination of Color and Local Patterns as a Feature Vector for CBIR Internatonal Journal of Computer Applcatons (975 8887) Volume 99 No.1, August 214 Combnaton of Color and Local Patterns as a Feature Vector for CBIR L.Koteswara Rao Asst.Professor, Dept of ECE Faculty

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

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

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

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

Performance Evaluation of Information Retrieval Systems

Performance Evaluation of Information Retrieval Systems Why System Evaluaton? Performance Evaluaton of Informaton Retreval Systems Many sldes n ths secton are adapted from Prof. Joydeep Ghosh (UT ECE) who n turn adapted them from Prof. Dk Lee (Unv. of Scence

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

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

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

HIGH-LEVEL SEMANTICS OF IMAGES IN WEB DOCUMENTS USING WEIGHTED TAGS AND STRENGTH MATRIX

HIGH-LEVEL SEMANTICS OF IMAGES IN WEB DOCUMENTS USING WEIGHTED TAGS AND STRENGTH MATRIX HIGH-LEVEL SEMANTICS OF IMAGES IN WEB DOCUMENTS USING WEIGHTED TAGS AND STRENGTH MATRIX P.Shanmugavadvu 1, P.Sumathy 2, A.Vadvel 3 12 Department of Computer Scence and Applcatons, Gandhgram Rural Insttute,

More information

Module Management Tool in Software Development Organizations

Module Management Tool in Software Development Organizations Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,

More information

Learning-Based Top-N Selection Query Evaluation over Relational Databases

Learning-Based Top-N Selection Query Evaluation over Relational Databases Learnng-Based Top-N Selecton Query Evaluaton over Relatonal Databases Lang Zhu *, Wey Meng ** * School of Mathematcs and Computer Scence, Hebe Unversty, Baodng, Hebe 071002, Chna, zhu@mal.hbu.edu.cn **

More information

Face Detection with Deep Learning

Face Detection with Deep Learning Face Detecton wth Deep Learnng Yu Shen Yus122@ucsd.edu A13227146 Kuan-We Chen kuc010@ucsd.edu A99045121 Yzhou Hao y3hao@ucsd.edu A98017773 Mn Hsuan Wu mhwu@ucsd.edu A92424998 Abstract The project here

More information

Real-time Motion Capture System Using One Video Camera Based on Color and Edge Distribution

Real-time Motion Capture System Using One Video Camera Based on Color and Edge Distribution Real-tme Moton Capture System Usng One Vdeo Camera Based on Color and Edge Dstrbuton YOSHIAKI AKAZAWA, YOSHIHIRO OKADA, AND KOICHI NIIJIMA Graduate School of Informaton Scence and Electrcal Engneerng,

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

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

Brushlet Features for Texture Image Retrieval

Brushlet Features for Texture Image Retrieval DICTA00: Dgtal Image Computng Technques and Applcatons, 1 January 00, Melbourne, Australa 1 Brushlet Features for Texture Image Retreval Chbao Chen and Kap Luk Chan Informaton System Research Lab, School

More 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

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

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

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

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 New Approach For the Ranking of Fuzzy Sets With Different Heights

A New Approach For the Ranking of Fuzzy Sets With Different Heights New pproach For the ankng of Fuzzy Sets Wth Dfferent Heghts Pushpnder Sngh School of Mathematcs Computer pplcatons Thapar Unversty, Patala-7 00 Inda pushpndersnl@gmalcom STCT ankng of fuzzy sets plays

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

KOHONEN'S SELF ORGANIZING NETWORKS WITH "CONSCIENCE"

KOHONEN'S SELF ORGANIZING NETWORKS WITH CONSCIENCE Kohonen's Self Organzng Maps and ther use n Interpretaton, Dr. M. Turhan (Tury) Taner, Rock Sold Images Page: 1 KOHONEN'S SELF ORGANIZING NETWORKS WITH "CONSCIENCE" By: Dr. M. Turhan (Tury) Taner, Rock

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

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

KIDS Lab at ImageCLEF 2012 Personal Photo Retrieval

KIDS Lab at ImageCLEF 2012 Personal Photo Retrieval KD Lab at mageclef 2012 Personal Photo Retreval Cha-We Ku, Been-Chan Chen, Guan-Bn Chen, L-J Gaou, Rong-ng Huang, and ao-en Wang Knowledge, nformaton, and Database ystem Laboratory Department of Computer

More information

Novel Pattern-based Fingerprint Recognition Technique Using 2D Wavelet Decomposition

Novel Pattern-based Fingerprint Recognition Technique Using 2D Wavelet Decomposition Mathematcal Methods for Informaton Scence and Economcs Novel Pattern-based Fngerprnt Recognton Technque Usng D Wavelet Decomposton TUDOR BARBU Insttute of Computer Scence of the Romanan Academy T. Codrescu,,

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

UB at GeoCLEF Department of Geography Abstract

UB at GeoCLEF Department of Geography   Abstract UB at GeoCLEF 2006 Mguel E. Ruz (1), Stuart Shapro (2), June Abbas (1), Slva B. Southwck (1) and Davd Mark (3) State Unversty of New York at Buffalo (1) Department of Lbrary and Informaton Studes (2) Department

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

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

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between

More 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

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

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

Clustering Algorithm of Similarity Segmentation based on Point Sorting

Clustering Algorithm of Similarity Segmentation based on Point Sorting Internatonal onference on Logstcs Engneerng, Management and omputer Scence (LEMS 2015) lusterng Algorthm of Smlarty Segmentaton based on Pont Sortng Hanbng L, Yan Wang*, Lan Huang, Mngda L, Yng Sun, Hanyuan

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

Data Modelling and. Multimedia. Databases M. Multimedia. Information Retrieval Part II. Outline

Data Modelling and. Multimedia. Databases M. Multimedia. Information Retrieval Part II. Outline ALMA MATER STUDIORUM - UNIVERSITÀ DI BOLOGNA Data Modellng and Multmeda Databases M Internatonal Second cycle degree programme (LM) n Dgtal Humantes and Dgtal Knowledge (DHDK) Unversty of Bologna Multmeda

More information

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

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

Robust Mean Shift Tracking with Corrected Background-Weighted Histogram

Robust Mean Shift Tracking with Corrected Background-Weighted Histogram Robust Mean Shft Trackng wth Corrected Background-Weghted Hstogram Jfeng Nng, Le Zhang, Davd Zhang and Chengke Wu Abstract: The background-weghted hstogram (BWH) algorthm proposed n [] attempts to reduce

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 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

Pruning Training Corpus to Speedup Text Classification 1

Pruning Training Corpus to Speedup Text Classification 1 Prunng Tranng Corpus to Speedup Text Classfcaton Jhong Guan and Shugeng Zhou School of Computer Scence, Wuhan Unversty, Wuhan, 430079, Chna hguan@wtusm.edu.cn State Key Lab of Software Engneerng, Wuhan

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

SRBIR: Semantic Region Based Image Retrieval by Extracting the Dominant Region and Semantic Learning

SRBIR: Semantic Region Based Image Retrieval by Extracting the Dominant Region and Semantic Learning Journal of Computer Scence 7 (3): 400-408, 2011 ISSN 1549-3636 2011 Scence Publcatons SRBIR: Semantc Regon Based Image Retreval by Extractng the Domnant Regon and Semantc Learnng 1 I. Felc Raam and 2 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

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

Vol. 5, No. 3 March 2014 ISSN Journal of Emerging Trends in Computing and Information Sciences CIS Journal. All rights reserved.

Vol. 5, No. 3 March 2014 ISSN Journal of Emerging Trends in Computing and Information Sciences CIS Journal. All rights reserved. Journal of Emergng Trends n Computng and Informaton Scences 009-03 CIS Journal. All rghts reserved. http://www.csjournal.org Unhealthy Detecton n Lvestock Texture Images usng Subsampled Contourlet Transform

More information

Enhanced Watermarking Technique for Color Images using Visual Cryptography

Enhanced Watermarking Technique for Color Images using Visual Cryptography Informaton Assurance and Securty Letters 1 (2010) 024-028 Enhanced Watermarkng Technque for Color Images usng Vsual Cryptography Enas F. Al rawashdeh 1, Rawan I.Zaghloul 2 1 Balqa Appled Unversty, MIS

More information

Hermite Splines in Lie Groups as Products of Geodesics

Hermite Splines in Lie Groups as Products of Geodesics Hermte Splnes n Le Groups as Products of Geodescs Ethan Eade Updated May 28, 2017 1 Introducton 1.1 Goal Ths document defnes a curve n the Le group G parametrzed by tme and by structural parameters n the

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

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

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

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

Efficient Content Representation in MPEG Video Databases

Efficient Content Representation in MPEG Video Databases Effcent Content Representaton n MPEG Vdeo Databases Yanns S. Avrths, Nkolaos D. Doulams, Anastasos D. Doulams and Stefanos D. Kollas Department of Electrcal and Computer Engneerng Natonal Techncal Unversty

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

Unsupervised Learning and Clustering

Unsupervised Learning and Clustering Unsupervsed Learnng and Clusterng Supervsed vs. Unsupervsed Learnng Up to now we consdered supervsed learnng scenaro, where we are gven 1. samples 1,, n 2. class labels for all samples 1,, n Ths s also

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

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces Range mages For many structured lght scanners, the range data forms a hghly regular pattern known as a range mage. he samplng pattern s determned by the specfc scanner. Range mage regstraton 1 Examples

More information

Fitting: Deformable contours April 26 th, 2018

Fitting: Deformable contours April 26 th, 2018 4/6/08 Fttng: Deformable contours Aprl 6 th, 08 Yong Jae Lee UC Davs Recap so far: Groupng and Fttng Goal: move from array of pxel values (or flter outputs) to a collecton of regons, objects, and shapes.

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

Palmprint Feature Extraction Using 2-D Gabor Filters

Palmprint Feature Extraction Using 2-D Gabor Filters Palmprnt Feature Extracton Usng 2-D Gabor Flters Wa Kn Kong Davd Zhang and Wenxn L Bometrcs Research Centre Department of Computng The Hong Kong Polytechnc Unversty Kowloon Hong Kong Correspondng author:

More information

3D Face Reconstruction With Local Feature Refinement. Abstract

3D Face Reconstruction With Local Feature Refinement. Abstract , pp.6-74 http://dx.do.org/0.457/jmue.04.9.8.06 3D Face Reconstructon Wth Local Feature Refnement Rudy Adpranata, Kartka Gunad and Wendy Gunawan 3, formatcs Department, Petra Chrstan Unversty, Surabaya,

More information