variable quality image compression system based on spiht
DESCRIPTION
Variable quality image compression system based on SPIHT. Authors: A. Jarvi, J. Lehtien, O. Nevalainen Source: Signal Processing: Image Communication 14(1999) 683-696 Date: 11 September 1999 Speaker: Ming-Ni Wu. SPIHT. - PowerPoint PPT PresentationTRANSCRIPT
Variable quality image compression system based on
SPIHT
Authors: A. Jarvi, J. Lehtien, O. Nevalainen
Source: Signal Processing: Image Communication 14(1999) 683-696
Date: 11 September 1999
Speaker: Ming-Ni Wu
SPIHT
• Set Partitioning In Hierarchical Trees by Said and Pearlman 1996
• a wavelet-based image compressing method
• to provide better performance then original EZW(Embedded Zerotree Wavelet)
• a bitplane coding method
SPIHT
n: the number of significant bits in the coefficient table having the largest absolute value
H: set of coordinates of all spatial orientation tree roots
LSP: List of Significant pixels
LIP: List of insignificant pixels
LIS: List of insignificant sets
type A: set of coordinates of all descendants of the node
type B: set of coordinates of all descendants of the node, excluding offspring of the node
vqSPIHT
• to provide better quality for some regions or details
• implementation with matrix
vqSPIHT
: a percentage of the wanted final output file size
n: the number of significant bits in the coefficient table having the largest absolute value
ROI: Region of Interest
LUT: lookup table, to define whether a coefficient(i,j) contribute to any ROI or not
vqSPIHT
PSM: point significance matrix11:significant00: insignificant01:unknown
SPLM:sub-pyramid list matrixtype A: set of coordinates of all descendants
of the node type B: set of coordinates of all descendants
of the node, excluding offspring of the node
Test Results
• The comic test image of size 420*480 with 8bpp
Test Results
• The mammogram test image of size 2185*2925 with 12bpp
Test Results
• Practical memory requirements of the implementation for 12MB mammogram– SPIHT : 120MB– vqSPIHT: 50MB
Conclusions
• to encode different parts of an image with different bit-rates depending on their importance
• to simplify the implementation of SPIHT• to reduce working storage requirements• Combine the sending of significant
coefficients to sorting phase to avoid the problem of transmission interrupting