On strange images with application to lossy image compression

Single and three-channel images are widely used in numerous applications. Due to the increasing volume of such data, they must be compressed where lossy compression offers more opportunities. Usually, it is supposed that, for a given image, larger ratio leads worse quality image according all metrics. This true most practical cases. However, has been found recently that called “strange” which rate-distortion curve like dependence peak signal-to-noise on factor or quantization step, behaves non-monotonously. might cause problems images. Thus, basic subject this paper factors determine phenomenon. The main among them artificial origin an possible presence large homogeneous regions, specific behavior histograms. goal consider explain peculiarities strange tasks provide definitions check whether non-monotonicity curves occurs different coders One task put ideas methodology forward further studies intended detect before their compression. result non-monotonous can observed same several metrics coders. means not coder but properties probability being strange. Moreover, both grayscale color strange, natural scene depends than number channels. In particular, percentage pixels belong regions entropy play important role. As conclusions, we outline directions future research first order, relate analysis databases establish parameters show considered as