Character Decompositions

Several decomposition formats exist for image data: principal components, independent components, non-negative components, etc. These decompositions have been applied mainly to natural image data. In this paper, we study the above decompositions for hand-written devanagari character data. We propose a new measure – spatial entropy – for characterizing datasets. Datasets with high spatial entropy are likely to give rise to local features.

[1]  Alex Pentland,et al.  Face recognition using eigenfaces , 1991, Proceedings. 1991 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[2]  Susan T. Dumais,et al.  Using latent semantic analysis to improve information retrieval , 1988, CHI 1988.

[3]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[4]  Stan Z. Li,et al.  Learning spatially localized, parts-based representation , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[5]  A. J. Bell,et al.  A Unifying Information-Theoretic Framework for Independent Component Analysis , 2000 .

[6]  H. Sebastian Seung,et al.  Algorithms for Non-negative Matrix Factorization , 2000, NIPS.