Image Processing: Lossy Compression by Color Quantization and GCT Modeling

A Study on Image Corruption and Noise in Lossy Compression

Abstract

Corruption may come in various forms such as motion, blur and camera resolution problem. In addition to these blurring effects, noise always corrupts any recorded image. Noise may be introduced by the medium through which the image is produced (random absorption or scatter effects), by the recording medium (sensor noise), by measurement errors due to the limited accuracy of the recording system, and by quantization of the data for digital storage.