Document Type
Thesis
Date of Award
5-31-1991
Degree Name
Master of Science in Electrical Engineering - (M.S.)
Department
Electrical and Computer Engineering
First Advisor
Ali N. Akansu
Second Advisor
Yeheskel Bar-Ness
Third Advisor
Zoran Siveski
Abstract
Subband coding of still image frames using Binomial PR-QMF has been presented in this thesis. Simulation results have shown that the performance of subband coding using a low complexity coder is practically the same as the performance of the industry standard (8 x 8) DCT based image coder.
A low bit rate adaptive video coding technique is also introduced in this thesis. The redundancy within adjacent video frames is exploited by motion compensated interframe prediction. The Motion Compensated Frame Difference (MCFD) signals are filtered by employing Binomial PR-QMF structure into four subbands. Then, the subbands are quantized using an efficient Motion Based Adaptive Vector Quantization (MBAVQ) algorithm. Here, the adaptation scheme is based on block motion vectors rather than local signal energy which was used in earlier works of several researchers. The new technique results in a reduction in bit rate by nearly (40%) due to the drop of the extra bits used for local variances. Moreover, for the video test sequances considered, MBAVQ method gives superior SNR results over local Variance Based Adaptive Vector Quantiztion (VBAVQ) scheme especially for high motion frames.
Recommended Citation
Mutlag, Hosam Fawzi, "A comparative study of image coding techniques : filter banks vs. discrete cosine transform" (1991). Theses. 2562.
https://digitalcommons.njit.edu/theses/2562
