Image Feature Extraction Based on an Extended Non-negative Sparse Coding Neural Network Model

This paper proposes an extended non-negative sparse coding (NNSC) neural network model for natural image feature extraction. The advantage for our model is to be able to ensure to converge to the basis vectors, which can respond well to the edge of the original images. Using the criteria of objective fidelity and the negative entropy, the validity of image feature extraction is testified. Furthermore, compared with independent component analysis (ICA) technique, the experimental results show that the quality of reconstructed images obtained by our method outperforms the ICA method.

[1]  Erkki Oja,et al.  Image feature extraction by sparse coding and independent component analysis , 1998, Proceedings. Fourteenth International Conference on Pattern Recognition (Cat. No.98EX170).

[2]  Patrik O. Hoyer,et al.  Non-negative sparse coding , 2002, Proceedings of the 12th IEEE Workshop on Neural Networks for Signal Processing.

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

[4]  Terrence J. Sejnowski,et al.  The “independent components” of natural scenes are edge filters , 1997, Vision Research.

[5]  David J. Field,et al.  Emergence of simple-cell receptive field properties by learning a sparse code for natural images , 1996, Nature.