Complex Algorithm of Image Wavelet Compression: Distortion Evaluation in the Light of Trade of Contour Separation and Compression Ratio

Abstract

The paper presents the results of estimating image distortions inherent for compression algorithms based on the complex implementation of wavelet coding taking into account the tradeoff between the degree of compression and the transmission of texture. Estimations of the level of distortion of image contours are given depending on the detail and texture characteristics on large areas of the image with a smooth brightness variation. The choice of test images containing low-contrast textures with different levels of medium brightness is justified. As a measure of the distortion between the original and the decoded image, the ratio of the peak value of the signal to the average noise value at the boundary of the contours of objects in the image is selected. The boundaries were detected by a gradient method. Compression factors are calculated for varying the thresholds of the restriction of the spectral components at different levels of the wavelet decomposition. A variant of the frequency-dependent restriction of the spectral components is proposed, which is determined for each of the subbands. Appropriate values of the compression ratio for the coding method based on the prediction of bit planes for a given set of images are presented.