Wavelet compression of ultrasound video streams for teleradiology

Abstract__This paper builds a probability estimating nodel of image m e s ; i o n ratio in different scan swuences, The c?xperiments show that the probable compressiorr ratio can beusedto estimate the real cnwression ratio. An algorithm to search in optid sequence autmatically from a given quanti zation matrix has been found. Using this algorithm we obtain a result that the optimal sequence of Y-matrix is just the scan sequence ZIG-ZAG. INlRowcTIONThe image compression proceclure eould be devided into 3 stages: transform , quantization and run-length encodling, later coa_lpression or later encoding. So we &o-the general compression ratio R asThe transforin stage mainly prepares the data for follow-up quantization, run-length encoding and later compression. Soroetimes it CoMjeRSes theer preliminarily, but the colllpression ratio R1 has nothing to do with the scan sequence. The quantization and run -length encoding of the sub-block i s in progress with same scan sequence, Using a selected quantization matrix to quantizate a definite image, the compression ratio R21 is perfect independent of the scan sequence too.The run-length ending arranges the 0-7803-2916-3 20. 10. 1 coefficients in the 2-dimention transform domain matrix to a 1-dimention form, and encodes the succesive "0'. The run-length encoding compression ratio R22 is ObviousIy relative with the scan sequence.Because of R22 is lossless, an optimized scan sequence can make full use of the transf0r-a and quantization results to take larger R22 without any increment of image quality loss.Later compression reduces the entropy redundant furtherly. Since the optimized scan sequence makes distribution of the run-length codes more concentrative, it is of great advantage to enlarge entropy enco -ding ccH8pression ratio R3, UP to nowt the scan sequerlce widely used in ilaage corspression is the ZIG-WIG. It is also reeomended by the welllau>wn R E G .In the Qmtrix, used earlier period,the is0 -quantization level bands are perpendicular to the main diagonal line nf the quantization matrix. ZI G-ZAG scan sequence s h o w no difference with the non-decrease order of element values of the quantization matrix and looks clearly reasonable. But now in the widely used quantization aratrix, such as the Y-signal default quantization mtrix recorrrmended by the CCITf hereinafter called Y-matrix ) , ZIG-ZAG is perfect different with the non-decrease order of the eleroent values. Then whether and why it i s still the optimal Scan sequence needsatheoretical analysis and demonstrat ion. %is paper aim to build a probable model to estimate the image compression ratio of different scan sequence, to get an algorithm finding the optimal sequence automatically

show abstract

Wavelet compression of ultrasound video streams for teleradiology

Cited by 6 publications

References 4 publications

Wavelet Packet Compression of Medical Images

Wavelet Packet Compression of Medical Images

Diagnostic ultrasound and telemedicine utilization in the international space station

The scan sequence of image compression and its optimization

Contact Info

Product

Resources

About