Date of Award

Fall 1995

Document Type

Thesis

Degree Name

Master of Science in Electrical Engineering - (M.S.)

Department

Electrical and Computer Engineering

First Advisor

Ali N. Akansu

Second Advisor

Nirwan Ansari

Third Advisor

Zoran Siveski

Abstract

Image compression is an important and active area of signal processing. All popular image compression techniques consist of three stages: Image transformation, quantization (lossy compression only), and lossless coding (of quantized transform coefficients).

This thesis deals with a comparative study of several lossy image compression techniques. First, it reviews the well-known techniques of each stage. Starting with the first stage, the techniques of orthogonal block transformation and subband transform are described in detail. Then the quantization stage is described, followed by a brief review of the techniques for the third stage, lossless coding.

Then these different image compression techniques are simulated and their rate-distortion performance are compared with each other. The results show that two-band multiplierless PR-QMF bank based subband image codec outperforms other filter banks considered in this thesis. It is also shown that uniform quantizers with a "dead-zone" perform best. Also, the multiplierless PR-QMF bank outperforms the DCT based on uniform quantization, but underperforms the DCT based on uniform quantization with a "dead-zone".

Share

COinS