Method Local-Monotonic Determination Lengths of Binary Block Codes of Clustered Transformants

Modern services for providing various types of information are associated with the generation of vast volumes of traffic that need to be transmitted via telecommunication networks. A significant portion of this traffic is composed of video data streams. Certain requirements are set for the quality of such information, including timely delivery, accuracy, visual perception quality, and the quality of structural component transmission in monitoring objects. As a result, data compression methods are implemented in information processing systems. Several families and versions of standardized approaches to compression have already been developed. However, the need for analyzing video data generated by remote sensors is increasingly evident. This creates a contradiction between such metrics as the timeliness of video transmission and its accuracy (quality of visual perception of the restored video images). Thus, a pertinent scientific and practical challenge lies in further improving video data compression methods to mitigate the conflict between these key metrics. The greatest interest is in metrics like the level of compression and the distortions in the digitized video image descriptions (measured by the peak signal-to-noise ratio, or PSNR). However, these metrics tend to exhibit inverse dependency, meaning that increasing the compression level often leads to a decrease in video quality. Consequently, the need to address this contradiction has prompted the development of a range of different image compression methods. Nevertheless, a common drawback of existing compression methods is that specified compression levels are often achieved in image quality modes that exhibit artifacts, contour distortions, and degradation of fine details. Therefore, for the further advancement of compression methods, it is proposed to choose an approach that involves controlled distortion of video quality. One potential direction for this approach is to utilize preprocessing in the spectral domain of video image fragments. Hence, the aim of this study is to develop compression methods with controlled levels of quality distortion based on preprocessing in the spectral domain. This article presents the main stages in developing a method for locally monotonic code determination for binary block codes in a differential-normalized space of structural components of clustered transformants in the spectral-parametric description (SPOT) using uniform-length markers. The method is based on a system for marking binary block codes of SPOT components by establishing interval domains for the arguments of block coding functions. This approach ensures the required integrity level of restored video fragments by creating conditions for the mutually unambiguous transformation of code formation processes. A comparative assessment of the compression ratio for the developed and existing methods revealed that, for significant SPOT transformants with quantization parameters forming PSNR levels from 27 to 37 dB, the developed method demonstrates an advantage in the compression ratio. Depending on the PSNR level, the gain over the methods used for comparison ranges from 14% to 21%.