Wavelet Identical Learning Neural Network–Based Image (JPG-JPEG) compression

Abstract

The best image quality at a given bit-rate (or compression rate) is the main goal of image compression to create much smaller files that use less space to store and less time to send. This demand is to increase the bandwidth to all users. The quality reduction achieved by manipulation of the bit stream or file without decompression and re-compression refers to scalability. In this paper the names for scalability are progressive coding or embedded bit streams, especially with a lossless compression found in a form of coarse-to-fine pixel scans. A fast scheme with a wavelet neural network fast forward multi- layer perceptron is used to recognize the quality or layer and resolution component progressive, (the highest possible compression rate without visual loss of image quality). Hence, the bit stream successively refines the reconstructed image by encoding the lower image resolution, then encoding the difference to high resolution. By encoding both grey and color pixels. With lossy compression of the image the quality of the compression procedure is measured by the Peak signal-to-noise ratio. In this paper a high resolution digital camera (DSC-W-100, 8.1 Mega Pixels) with a (2448x3264x3) or the total grand is (23970816) elements using (23970816) bytes for my baby photo. The language used is (Mat-Lab languages).