QUANTIFICATION CONTROL METHOD AND SYSTEM BASED ON FrDCT TRANSFORMATION MATRIX IN REAL-TIME IMAGE COMPRESSION SYSTEMS

Abstract

The invention relates to an image compression method and system for processing Fractional Order Discrete Cosine Transform (FrDCT) matrices used in image compression systems with low computational cost. Within the scope of the invention, a control parameter is created by processing the coefficient densities of different sub-regions of the FrDCT transform matrix, and this control parameter is used to control the quantization operation of the image compression engine. Thus, quantization control based on the transform matrix is ​​performed without re-running the entire JPEG compression chain. Thanks to this invention, the computational load, complexity, and memory access intensity of the image compression system can be reduced. The invention is particularly applicable to FPGA-based image processing systems, DSP-based codec architectures, embedded image compression systems, and real-time image processing applications.