An investigation of vector quantization-based methods for image compression in color printers

Advancements in image fidelity and color reproduction systems have made it possible to print images with nearly photographic quality. The abundance of digital image processing techniques to enhance or manipulate the image data is making color printing a viable alternative to photography. Unfortunately, an increase in the image resolution and depth of color leads to a severe stress on memory and network resources. Image compression helps in reducing this load. The algorithms employed should take into account the printer's limited computing and memory resources.; The main goals of the compression system should be high fidelity and ease of decoding. This thesis looks at the application of vector quantization based techniques in providing an asymmetric encoding and decoding system to address this problem. The computational advantages of using vector quantization as opposed to JPEG are brought out. A subjective evaluation test is undertaken to justify the use of vector quantization.