Wavelet Image Compression Method Combined with the GPCA

In the most common standard coding of the image, coder has greater complexity than the decoder, usually 5-10 times higher. In wireless sensor networks is the opposite, the coder implemented in a battery feed device has very limited resources, while decoder is typically implemented in a powerful computer. The paper focuses on a combination of wavelet technique with algebraic GPCA method, realizing an improvement of traditional wavelet, providing a suitable compression of multimedial information, without reducing its quality. Simulations in MATLAB, provide satisfactory compression levels. The proposed method, unlike other existing methods, takes into account the multi-modal data structure, except correlation in 1 D and 2 D, it takes in consideration also the correlation between color channels. The method gives a significant improvement in the PSNR values that reach an average of 15% compared with classic wavelet compression.