International Journal of Reviews in Computing © 2009-2010 IJRIC & LLS. All rights reserved. ISSN: 2076-3328 www.ijric.org IJRIC E-ISSN: 2076-3336 AFFINE INVARIANT DESCRIPTORS AND RECOGNIZING OF 3D OBJECTS USING STATISTICS METHODS 1 M. ELHACHLOUFI , 1A. EL OIRRAK, 2D. ABOUTAJDINE , 1M.N. KADDIOUI 1 Faculty Semlalia , Department of Informatics, Marrakech, Morocco 2 Faculty of Science, LEESA-GSCM, BP 1014, RABAT E-mail: elhachloufi@yahoo.fr , oirrek@yahoo.fr , aboutaj@fsr.ac.ma, kaddioui@ucam.ac.ma ABSTRACT The increasing number of objects 3D available on the Internet or in specialized databases which require the establishment of methods to develop description and recognition techniques to access intelligently to the contents of these objects. In this context, our work whose objective is to present the methods of invariant description [1,2,3] and recognition of 3D objects based on statistical methods: analysis of data (AD) and the principal component analysis (PCA). The objective of this studies is to determine an invariant description [4,5] of a 3D object using a coefficients vector of canonical correlations from the data analysis. This vector is invariant against affine transformation of the 3D object and recognize the object (s) of a database (3D objects) similar(s) to a given object (query object) using the descriptor vectors extracted from these 3D objects by the principal component analysis.The 3D objects of this database are transformations of 3D objects by one element of the overall transformation. The set of transformations considered in this work is the general affine group. The measure of similarity between two objects is achieved by a similarity function using the Euclidean distance. Keywords: Invariants Descriptors, Recognizing, 3D objects, Multiple Regression, Principal Component Analysis, Affine Transformation. 1. INTRODUCTION transformations and algebraic invariants [10], which provide global descriptors, which are expressed in terms of moments of different orders. With the advent of the Internet, exchanges and the acquisition of information, description and recognition of 3D objects have been as extensive and have become very important in several domains. On the other hand, the size of 3D objects used on the Internet and in computer systems has become enormous, particularly due to the rapid advancement technology acquisition and storage which require the establishment of methods to develop description and recognition techniques to access intelligently to the contents of these objects. In fact, several approaches are used: in terms of statistical approaches, the statistical shape descriptors for recognition in general consist either of calculating various statistical moments [6] [7] and [8], or of estimating the distribution of the measurement of a given geometric primitive, when either deterministic or random. Among the approaches by statistical distribution, we mention the specter of 3D shape (SF3D) [9] which is invariant to geometric For structural approaches, approaches representative of the object segmentation in 3D plot of land and performances by adjacency graph are presented in [11] and [12]. Similarly, Tangelder and al [13] have developed an approach based on representations by interest points. In transform approaches a very rich literature emphasizes any interest in approaches based transform Haugh [14], [15] and [16] which consists in detecting different varieties of dimension (n-1) immersed in the space. In the same vein, this work interested to define new methods for description and recognition of 3D objects using statistics methods. 93 International Journal of Reviews in Computing IJRIC © 2009-2010 IJRIC & LLS. All rights reserved. ISSN: 2076-3328 2. www.ijric.org 12V Corr (V 1 ,V 2 ) 0 , METHOD 1 : AN AFFINE DESCRIPTION INVARIANT OF 3D OBJECTS BY THE CANONICAL CORRELATION COEFFICIENTS linear 2 uncorrelated W 1 k W1 to 2 2 and 2 W are correlated as possible, i.e.: maximizes 2 2 the quantity available 2 Corr (V , W ) . And so on... Consider 3D object represented by a set of denoted M and W 2Y i.e.: 12 Corr (W , W ) 0 , witch V 2.1 Representation of the 3D objects points E-ISSN: 2076-3336 Pi i 1.......n , The canonical analysis product p of pairs of ( xi , y j , zi ) , arranged in a matrix X . Under the action of an affine transformation, the coordinates ( x, y, z ) are transformed into other coordinates ( x, y, z ) by variables where s = 1... p. The variables V are an orthogonal basis of the space generated the following procedure: f : 3 3 X ( x(t ), y (t ), z (t )) f ( X ) Y ( x(t ), y (t ), z (t )) Y A X ( x(t ), y (t ), z (t )) B particularly the first of them, reflect the linear connections between two groups of initial where Pi With s 3 A (aij )i , j 1,2,3 invertible by X 3 matrix V 1 vectors. , W 1 which j of variables X j (X j (decreasing) are called ,W s s 1,..., p ( s ) s 1,....., p is j column of X) and of X and Y of canonical from these couples. then the correlations k variables Figure 4 shows the values of canonical correlation coefficients are all equal to 1. Figures 3 shows that the canonical variables of origin objects and its transformed are the same which leads us to conclude that Y is an affine transformation of X according to the procedure of canonical analysis. Y k ( Y k is kth column), such that V 1 1 and W are correlated as possible, i.e: that maximizes the quantity available 1 Corr (V 1 ,W 1 ) . Then we search where s coefficients th W 1Y a normalized linear combination of 1 k Corr (V k ,W k ) V V 1 X is a normalized linear combination 1 s variables. The variables V are W are called canonical variables. Their successive correlations We consider two 3D objects X and Y related by an affine transformation (figure 1 and 2). The calculation of canonical correlation coefficients from these objects requires computing the first a pairs of canonical variables The principle of this analysis is to search at the first a couple of variables , 2.3 Results and evaluation t two s s over an affine transformation and are all equal to 1 in this case. X X 1 , X 2 ,........, X p and are s canonical correlation coefficients (or canonical correlations). The canonical correlation ( ) coefficients s s 1,....., p are invariant quantities . Y (Y1 , Y2 ,........., Yq )t k 2.2 The canonical analysis of two vectors: X and Y Let s . The variables W is e an orthogonal space generated by Y . The couples V , W associated with the infinite, and B is a vector translation in j the normed pair V 2 2 X j as combination of X j being V 2 a uncorrelated to V ,W 2 linear 1 Figure 1: 3D Objet origin , i.e.: 94 Figure 2: 3D Objet transformed International Journal of Reviews in Computing IJRIC © 2009-2010 IJRIC & LLS. All rights reserved. ISSN: 2076-3328 www.ijric.org E-ISSN: 2076-3336 ( , u , v ) is a vector characteristic of X , so we write : X ( , u , v ) Figure 3: Representation of the canonical variables of 3D object original as function of the 3D object transformed (1) 3.2 Representation of 3D objects Figure 4: Representation of the canonical correlation coefficients Consider a 3D object represented by a set of Pi i 1....... n points denoted with Pi t ( xi , y j , zi ) , arranged in a matrix X , 3. X t P1 , P2 ,..., Pn where X t and P t are the matrix and vector transposed X and P . i.e: METHOD 2: RECOGNITION 3D OBJECT PRINCIPAL COMPONENT ANALYSIS NORMALIZED Under the action of an affine transformation, the coordinates of X are transformed into other coordinates of Y by the following procedure: Y X where and are scalar. 3.1 Principal component analysis The principal component analysis is a method factorial analysis of multidimensional data [18]. It determines a decomposition of a X random vector with uncorrelated components, orthogonal and adjusting to better distribution of X . In this sense the components are called principal components and are arranged in descending order according to their degree of adjustment. The calculation of normalized principal components of the vector is carried out initially by calculating the covariance matrix as follows: V Let f function defined as follows: a f f (a ) where vi Xui i and are the have : Y Y Then f is called an invariant function. Remarque: (a)- If X Y then X Y . (b)We suppose: Y X then X f f ( X ) f (Y ) Y f . According to (a) 2- X u u u u1 , u2 .....u p and u are eigenvalues and eigenvectors associated V . The eigenvalues and eigenvectors v and a2 X ( X ) Y var( X ) (X X ) (X X ) X X f (X ) X X 2 var( X ) f (Y ) 1- Det (V I ) 0 1 , 2 ..... p (2) mean and variance X . We Then we passes to extract eigenvalues and eigenvectors associated to V by the following process t a where a is un 3D object , a and 1 t X X , Where X t transposed of X . n associated V are aa we obtain : v (v1 ,..., vn ) X Y f f , thus (.) f is an invariant against affine transformation. (c)- According to (2) we concluded that: n is the number of line X X f X X X ( , u , v ) of X . The reconstruction of X from vector ( , u , v ) is doing as follows If X i ui v , we say that the vector X f i ui vit f (3) is a vector characteristic of X f , i.e : t i i i 95 (4) International Journal of Reviews in Computing IJRIC © 2009-2010 IJRIC & LLS. All rights reserved. ISSN: 2076-3328 then www.ijric.org X f ( X f , X , X ) is a descriptor of descriptor, we calculate the error resulting from the difference between this vector and that of the object (query object) to recognize. We recognize the object that produced the error is almost zero. This operation is illustrated in figure 6. X in the sense that the invariance check is only relative to the first component X and the f reconstitution X requires the use of and X E-ISSN: 2076-3336 X , X f as shown in equation (4). Unkhown object 3.3 Principle of description and recognition proposed method Extraction of feature vector Y associate Calculation of Y Y f f (Y ) for this object 3.3.1 Step 1: learning system Evaluation of It can be decomposed into two phases: extraction of vector descriptor and recognition. The role of the first phase is to associate each object to learn the vector descriptor Y Database X f ( X f , X , X ) . This vector will be used erri X i Y later in the recognition system. Registration vector descriptors are Phase 2. The diagram of learning system is shown in figure 5. If erri 0 then this object is recognized as an affine transformation of Input object the object Xi , we write X f f (X ) X X X Evaluate If not , this object is not recognized as an affine transformation of the object Xi X Extraction of feature vector X X i X fi X i X i Figure 6 : Diagram of recogntion system f associate for this object X E( X ), X var( X ) 3.4 Results and evaluation Consider two 3D objects X (object of a database) and Y (query object) related by an affine transformation ( figures 1 and 2). After extraction of vector descriptors X of X X f ( X f , X , X ) f Y f of Y we move to calculate the error vector err X Y (5) Database According to the graph of the error (figure 9) we found that err 0 , which verifies the invariant of this vectors as shown in figure 8 and figure 9. So Y is an affine transformation of X . Figure 5 : Learning System 3.3.2 Step 2 : Recognition system For each characteristic vector (first component) of the object database vector 96 International Journal of Reviews in Computing © 2009-2010 IJRIC & LLS. All rights reserved. ISSN: 2076-3328 www.ijric.org Figure 8 :descriptor of 3D objet origin using PCA Figure 8: descriptor of 3D objet transformed using PCA In the second method, the vector descriptor is formed by the characteristic vector extracted using PCA from the object origin and its transformed, this vector is invariant against an affine transformation. The recognition is doing by measuring the similarity between the descriptor vectors extracted from the two objects: the request object and those belonging to a database of objects using the metric euclidean distance as described above for each method. Experimental results show the validity and utility of these methods. err X Y 4. COMPARISON OF RESULTS According to figures below we can see that the error1 - corresponds to the difference between the invariants of the object origin and its transformation by an affine transformationcalculated by our methods (figures 9) is smaller than those obtained by the methods of Fourier transform and Moments (figures 10 and 11). In addition, the computing time of our method is less than the method of Fourier transform and Moments. This leads us to conclude that our method applied to these objects is more efficient than Fourier transform and Moments, in term of error. REFERENCES [1] [2] [3] [4] [5] Figure 10: error1 Figure 11: error1 representation by Fourier transform representation by Moments E-ISSN: 2076-3336 data (AD) and the principal component analysis (PCA). The first method consists of calculating the canonical correlations coefficients vector from the canonical variables couples of the object origin and it’s transformed. The components of this vector are all equal to 1 and the canonical variables values are equal which leads us to conclude that the first object (origin) is an affine transformation of the second (its transformed). Figure 9: Representation of error : IJRIC [6] 5. CONCLUSION In this work we presented two methods of invariant description and recognition of 3D objects based on statistical methods: analysis of [7] 97 A. El Oirrak, M. Daoudi, D. Aboutajdine: Affine invariant descriptors for color images using Fourier series. Pattern Recognition Letters 24(9-10): 1339-1348 (2003) A. El Oirrak, M. Daoudi, D.Aboutajdine: Estimation of general 2D affine motion using Fourier descriptors. Pattern Recognition 35(1): 223-228 (2002) A.El Oirrak, M. Daoudi, D.Aboutajdine: Affine invariant descriptors using Fourier series. Pattern Recognition Letters 23(10): 1109-1118 (2002) M. Petrou and A. Kadyrov. Affine invariant features from the trace transform. IEEE Transactions on Pattern Analysis and Machine Intelligence, 26(1):30{44, 2004. A. Choksuriwong, H. Laurent & B. Emile. A Comparative Study of Objects Invariant Descriptor. ORASIS, 2005. T. Murao, "Descriptors of polyhedral data for 313- shape similarity search", Proposal P177, MPEG-7 Proposal Evaluation Meeting, UK, Fév. 1999. M. Elad, A. Tal, S. Ar, "Directed search in a 3D objects database using SVM", International Journal of Reviews in Computing © 2009-2010 IJRIC & LLS. All rights reserved. ISSN: 2076-3328 www.ijric.org Hewlett-Packard Research Report HPL-2000-20R1, 2000. [8] C. Zhang, T. Chen, "Efficient feature extraction for 2D/3D objects in mesh representation", Proc. of the International Conference on Image Processing (ICIP 2001), 2001. [9] T. Zaharia, F. Prêteux, "3D-shape-based retrieval within the MPEG-7 framework", Proc. SPIE Conf. On Nonlinear Image Processing and Pattern Analysis XII, Vol. 4304, pp. 133-145, San Jose, EtatsUnis, Janv. 2001. [10] Taubin et D.B. Cooper, "Object recognition based on moment (or algebraic) invariants", J.L. Mundy and A. Zisserman, editors, Geometric Invariants in Computer Vision, pp. 375397, MIT Press, 1992. [11] S. J. Dickinson, D. Metaxas, A. Pentland, "The role of model-based segmentation in the recovery of volumetric parts from range data", IEEE Trans. On PAMI, Vol. 19, No. 3, pp. 259-267, Mars 1997. [12] S. Dickinson, A. Pentland, S. Stevenson, "Viewpoint-invariant indexing for content- based image retrieval", Proc. IEEE Int. Workshop on Content-Based Access of Image and Video Database, pp. 20-30, Janv. 1998. [13] J.W.H. Tangelder, R.C. Veltkamp, "Polyhedral model retrieval using weighted point sets", Rapport technique no UU-CS-2002-019, Universitd de Utrecht, Pays-Bas, 2002. [14] P. V. C. Hough, "Method and means for recognizing complex patterns", U.S. Patent 3069 654, 1962. [15] D. H. Ballard, "Generalizing the Hough transform to detect arbitrary shapes," Pattern Recognition, Vol. 13, No. 2, pp. 111-122, 1981. [16] J. Illingworth et J. Kittler, "A survey of the Hough transform", Computer Vision, Graphics and Image Processing, Vol. 44, pp. 87-116, Oct. 1988. [17] Bouroche, J.M. et G. Saporta (2006), L’analyse de données, Presses Universitaires de France, Collection Que sais-je ?, 9ème édition, 128p. 2006 [18] Saporta, G. (2006), Probabilités, analyse des données et statistiques, Techniques, 2ème édition, 622 p, 2006 98 IJRIC E-ISSN: 2076-3336