Title :
A wavelet block chain technique for use in zerotree coding
Author :
Syed, Yasser F. ; Rao, K.R.
Author_Institution :
Dept. of Electr. Eng., Texas Univ., Arlington, TX, USA
Abstract :
This paper discusses a novel approach towards low bit rate image coding using wavelet block chains (WBC) to code wavelet coefficient values in the LL subband while simultaneously providing segmentation information to the decoder. This technique forms chains of similar valued wavelet blocks in low spatial frequency regions from wavelet coefficients representing the image at its coarsest resolution. Because of the frequent occurrence of these regions in the LL subband, many chains of wavelet blocks can be formed across an image frame. Improvement in compression is achieved by transmitting the common coefficient value and directional information necessary to reconstruct the chains at the decoder. In addition, the WBC provides segmentation information to the decoder without additional overhead. This technique is compatible with zero tree wavelet compression methods and its effectiveness is shown by improved results in low bit rate zerotree entropy (ZTE) coding.
Keywords :
discrete wavelet transforms; image coding; image representation; image segmentation; transform coding; tree data structures; LL subband coding; chain coding; coarsest resolution; decoder; directional information; image coding; low bit rate coding; low bit rate zerotree entropy coding; low spatial frequency regions; segmentation information; wavelet block chain technique; wavelet coefficients; zero tree wavelet compression; zerotree coding; Bit rate; Decoding; Entropy; Frequency; Image coding; Image reconstruction; Image resolution; Image segmentation; Spatial resolution; Wavelet coefficients;
Conference_Titel :
Signals, Systems & Computers, 1998. Conference Record of the Thirty-Second Asilomar Conference on
Conference_Location :
Pacific Grove, CA, USA
Print_ISBN :
0-7803-5148-7
DOI :
10.1109/ACSSC.1998.751530