Application of adaptive constructive neural networks to image compression

IEEE Transactions on Neural Networks - Tập 13 Số 5 - Trang 1112-1126 - 2002
L. Ma1, K. Khorasani1
1Department of Electrical and Computer Engineering, Concordia University, Montreal, QUE, Canada

Tóm tắt

The objective of the paper is the application of an adaptive constructive one-hidden-layer feedforward neural networks (OHL-FNNs) to image compression. Comparisons with fixed structure neural networks are performed to demonstrate and illustrate the training and the generalization capabilities of the proposed adaptive constructive networks. The influence of quantization effects as well as comparison with the baseline JPEG scheme are also investigated. It has been demonstrated through several experiments that very promising results are obtained as compared to presently available techniques in the literature.

Từ khóa

#Adaptive systems #Neural networks #Image coding #Cellular neural networks #Image storage #Pixel #Image reconstruction #Feedforward neural networks #Quantization #Transform coding

Tài liệu tham khảo

cun, 1990, optimal brain damage, Advances in Neural Information Processing, 598 10.1109/5.364461 fahlman, 1991, The cascade-correlation learning architecture jayant, 1984, Digital Coding of Waveforms Principles and Applications to Speech and Video 10.1016/S0923-5965(98)00041-1 10.1109/72.80236 10.1109/72.572102 10.1109/72.623214 10.1007/978-1-4615-3954-4 10.1109/7.481272 10.1117/1.601326 10.1002/ett.4460030612 10.1109/TENCON.1997.647340 10.1109/72.572092 cottrell, 1987, learning internal representations from gray-level images: an example of extensional programming, Proc 9th Annu Conf Cognitive Sci Soc, 461 10.1109/76.322995 10.1080/09540098908915647 10.1109/72.668891 10.1016/S0377-2217(97)00370-6 10.1007/978-1-4899-6950-7 10.1109/ICNN.1994.375045 10.1142/S012906579700015X 10.1109/81.662694 10.1117/1.600828 10.1016/0893-6080(96)00032-9