A Texture-Based Statistical Model for Face Detection

Texture is an important cue for detecting objects that undergo shape deformation, pose changes and variations in illumination. We propose a general statistical model which relies on texture for learning an object class from a set of example images. We use the class of human faces to test our ideas. Once the model learns the distribution of the face images, it can then be used to classify new images as faces or not. The distribution of images is captured by a set of feature vectors which have been shown to produce good texture detection and synthesis results 4]. Our statistical model uses these feature vectors for classiication and can handle larger variations in the appearance of faces than previous approaches. We estimate the distribution of feature vectors for an object class by clustering the data and then forming a mixture of gaussian model. The mixture model is further reened by determining which clusters are the most discriminative for the class and retaining only those clusters. After the model is learned, test images are classiied by computing the likelihood of their feature vectors with respect to the model. We present excellent results in applying our technique to face detection. Abstract Texture is an important cue for detecting objects that undergo shape deformation, pose changes and variations in illumination. We propose a general statistical model which relies on texture for learning an object class from a set of example images. We use the class of human faces to test our ideas. Once the model learns the distribution of the face images, it can then be used to classify new images as faces or not. The distribution of images is captured by a set of feature vectors which have been shown to produce good texture detection and synthesis results 4]. Our statistical model uses these feature vectors for classiication and can handle larger variations in the appearance of faces than previous approaches. We estimate the distribution of feature vectors for an object class by clustering the data and then forming a mixture of gaussian model. The mixture model is further reened by determining which clusters are the most discriminative for the class and retaining only those clusters. After the model is learned, test images are classiied by computing the likelihood of their feature vectors with respect to the model. We present excellent results in applying our technique to face detection.

[1]  Pavel Pudil,et al.  Introduction to Statistical Pattern Recognition , 2006 .

[2]  Michael Jones,et al.  Multidimensional Morphable Models: A Framework for Representing and Matching Object Classes , 2004, International Journal of Computer Vision.

[3]  Bernt Schiele,et al.  Recognition without Correspondence using Multidimensional Receptive Field Histograms , 2004, International Journal of Computer Vision.

[4]  Eero P. Simoncelli,et al.  Texture characterization via joint statistics of wavelet coefficient magnitudes , 1998, Proceedings 1998 International Conference on Image Processing. ICIP98 (Cat. No.98CB36269).

[5]  Takeo Kanade,et al.  Probabilistic modeling of local appearance and spatial relationships for object recognition , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).

[6]  Paul A. Viola,et al.  Texture recognition using a non-parametric multi-scale statistical model , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).

[7]  Tomaso A. Poggio,et al.  A general framework for object detection , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).

[8]  Song-Chun Zhu.  Filters, Random Fields and Maximum Entropy (FRAME): Towards a Unified Theory for Texture Modeling , 1998 .

[9]  Tomaso A. Poggio,et al.  Example-Based Learning for View-Based Human Face Detection , 1998, IEEE Trans. Pattern Anal. Mach. Intell..

[10]  Eero P. Simoncelli,et al.  Embedded wavelet image compression based on a joint probability model , 1997, Proceedings of International Conference on Image Processing.

[11]  Jeremy S. De Bonet,et al.  Multiresolution sampling procedure for analysis and synthesis of texture images , 1997, SIGGRAPH.

[12]  Norbert Krüger,et al.  Face recognition by elastic bunch graph matching , 1997, Proceedings of International Conference on Image Processing.

[13]  Kris Popat,et al.  Cluster-based probability model and its application to image and texture processing , 1997, IEEE Trans. Image Process..

[14]  Paul A. Viola.  Complex Feature Recognition: A Bayesian Approach for Learning to Recognize Objects , 1996 .

[15]  Alex Pentland,et al.  Bayesian face recognition using deformable intensity surfaces , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[16]  Takeo Kanade,et al.  Human Face Detection in Visual Scenes , 1995, NIPS.

[17]  William T. Freeman,et al.  Presented at: 2nd Annual IEEE International Conference on Image , 1995 .

[18]  James R. Bergen,et al.  Texture Analysis: Representation and Matching , 1995, ICIAP.

[19]  C. Burly,et al.  Face Localization via Shape Statistics , 1995 .

[20]  Tomaso Poggio,et al.  Example Based Image Analysis and Synthesis , 1993 .

[21]  Joachim M. Buhmann,et al.  Distortion Invariant Object Recognition in the Dynamic Link Architecture , 1993, IEEE Trans. Computers.