Abstract

It is a common task in image lossy compression to provide a reasonable trade-off between a desired compression ratio (as large as possible) and an appropriate (preset according to a chosen metric) quality. Recently, a Better Portable Graphics (BPG) coder has been proposed that might become a new standard. A quality parameter Q is used in it to vary image quality and compression ratio. However, a given Q for particular images to be compressed produces quite different compression ratios and quality measured by a given quality metric. To provide a given quality of compressed images characterized by a visual quality metric FSIMc, we propose and analyze a two-step algorithm based on preliminary obtained average rate/distortion curve and its linear approximation. For this algorithm, the second step that calculates the improved estimate of Q and employs it for final compression allows decreasing variance of the produced FSIMc by about one order of magnitude compared to the first compression step where Q is set according to the average rate-distortion curve. The algorithm is applied for a wide set of desired values of FSIMc. It is also tested for a set of images not used in obtaining the average rate/distortion curve.

Citations

Authors (4)