Authors : Sin Ting Lim and Nurulfajar Bin Abdul Manap
Abstract: Medical image contains diagnostically important regions that shall not be subjected to lossy compression. In order to increase compression rate for higher transmission and storage capability, a partial compression scheme based on ROI and non-ROI was employed. A manual segmentation technique to separate ROI and non-ROI for thousands of images are impractical, hence in this study an automated brain segmentation technique was developed to work with a PCA compression scheme. Non-ROI region will be compressed by PCA compression while ROI region will be preserved. The segmentation technique specifically tailored for brain segmentation has successfully separate ROI and non-ROI regions and results indicate that image quality is higher for image undergo the proposed model compared with image without ROI segmentation.
Sin Ting Lim and Nurulfajar Bin Abdul Manap, 2018. Automated ROI-Based Compression on Brain Images Using Principal Component Analysis. Journal of Engineering and Applied Sciences, 13: 5961-5970.