Region-based medical image compression in teleradiology
M. Moorthi, R. Amutha · International Journal of Telemedicine and Clinical Practices · 2015
Digital radiology has resulted in significant increase in use of digital medical images in the process of diagnosis. Lossless image compression is necessary to preserve the value of diagnostic medical images. This method will be applicable to mainly compressing information without any loss. More particularly, this paper aims to increase compression ratio and quality (PSNR) of medical images. Initially, MRI brain image is segmented using grow cut. This algorithm extracts the region of tumour (abnormal region) and non-region of tumour (normal region) parts. The region of tumour is selected using seed point which is selected by extracting run length features. Morphological processing is used for highlighting the tumour region. The integer wavelet transform is applied on both region of tumour and non-region of tumour parts. The abnormal region is compressed losslessly using arithmetic coding and normal region is compressed using lossy compression (i.e.) EZW coding. Finally, the compressed image of the region of tumour (abnormal region) is fused with compressed image of the normal region. This implementation is found to possess superior reconstruction properties and better compression ratio.