Abstract

Lossy compression plays a vital role in modern digital image processing for producing a high compression ratio. However, distortion is unavoidable, which affects further image processing and must be handled with care. Providing a desired visual quality is an efficient approach for reaching a trade-off between introduced distortions and compression ratio; it aims to control the visual quality of the decompressed images and make them not worse than the required by a user. This paper proposes an intelligent lossy compression method of providing a desired visual quality, which considers the complexity of various images. This characteristic is utilized to choose an appropriate average rate-distortion curve for an image to be compressed. Experiments have been conducted for Discrete Cosine Transform (DCT) based lossy compression coder, Peak Signal-Noise Ratio (PSNR) has been employed to evaluate the visual quality. The results show that our new method has the ability to provide a general improvement of accuracy, and the proposed algorithm for classifying image complexity by entropy calculation is simpler and faster than earlier proposed counterparts. In addition, it is possible to find “strange” images which produce the largest errors in providing a desired quality of compression.

Citations

Authors (5)