Image coding using self-supervised backpropagation neural network

Image data compression is implemented using a self-supervised backpropagation neural network. First, a backpropagation discrete cosine transform (BPDCT) is developed and then used for transform coding. Secondly, to alleviate edge distortion, classification techniques are applied to transform image coding. The classification technique is based on edge detection since the human visual system is more sensitive to edges. Simulation results show that the BPDCT works better for image coding than a truncated DCT. Classification techniques improve the performance of the BPDCT.<<ETX>>

[1]  P. Yip,et al.  Discrete Cosine Transform: Algorithms, Advantages, Applications , 1990 .

[2]  J. Gimlett,et al.  Use of "Activity" Classes in Adaptive Transform Image Coding , 1975, IEEE Trans. Commun..

[3]  P. J. Werbos,et al.  Backpropagation: past and future , 1988, IEEE 1988 International Conference on Neural Networks.

[4]  Yo-Sung Ho,et al.  Classified transform coding of images using vector quantization , 1989, International Conference on Acoustics, Speech, and Signal Processing,.

[5]  M. T. Manry,et al.  Analysis of non-Gaussian data using a neural network , 1989, International 1989 Joint Conference on Neural Networks.

[6]  Anil K. Jain.  Fundamentals of Digital Image Processing , 2018, Control of Color Imaging Systems.

[7]  PAUL J. WERBOS,et al.  Generalization of backpropagation with application to a recurrent gas market model , 1988, Neural Networks.

[8]  D. Levine.  Introduction to Neural and Cognitive Modeling , 2018 .

[9]  K. Ngan.  Adaptive transform coding of video signals , 1982 .

[10]  C. Smith,et al.  Adaptive Coding of Monochrome and Color Images , 1977, IEEE Trans. Commun..