Compressed Medical Image Transfer in Frequency Domain

International Journal of Image Processing

Haval Mohammed Sidqi¹  aree ali mohammed¹ 

[1] $$

关键词: Haar Wavelet;    Region of Interest Coding;    Adaptive Quantization;   

DOI  :  

来源: Computer Science Journals

【 摘 要 】

A common approach to the medical imagecompression algorithm begins by separating theregion of interests from the background of themedical images and then lossless and lossycompression schemes are applied on the ROI partand background respectively. The compressed files(ROI and background) are now transmitted throughdifferent media of communications (local host,Intranet and Internet) between the server and clients.In this work, a medical image transfer codingscheme based on lossless Haar wavelet transformsmethod is proposed. At first, the proposed scheme istested on Intranet (for both RoI and background) inorder to compare its results with Internet tests. Anadaptive quantization algorithm is used to apply onquasi lossless ROI wavelet coefficients while auniform quantization is used to apply on lossybackground wavelet coefficients. Finally, theretained quantization indices are entropy encodedwith an optimal variable coding algorithm. The testresults have indicated that the performance of theproposed MITC via Intranet is much better than viaInternet in terms of transferring time, while thequality of the reconstructed medical image remainsconstant despite the medium of communication. Forbest adopted parameters, a compressed medicalimage file (760 KB „³ 19.38 KB) is transmittedthrough Internet (bandwidth= 1024 kbps) withtransfer time = 0.156 s while the uncompressed fileis sent with transfer time = 6.192 s.

【 授权许可】

Unknown