Region based medical image compression using block-based PCA
Ravi Kiran, Chandrashekhar Kamargaonkar · 2016 International Conference on Computation of Power, Energy Information and Commuincation (ICCPEIC) · 2016
Medical image compression has received great attention attributable to its increasing need to decrease the image size while not compromising the diagnostically crucial medical data exhibited on the image. PCA algorithm may be used to help in image compression. In this paper, a system that investigates the effectiveness of the region-based compression using block based PCA was proposed. The block based PCA has 2 extended-PCA algorithms that manipulate the block data of the image are tested. The first algorithm is referred to as block-by-block PCA where general PCA algorithm is applied to every block of the image. In the next algorithm- the block-to-row PCA, all block data are initially concatenated into a row before the general PCA algorithm is therefore applied in the transformed matrix. In this work, the automated segmentation is employed to trace the specified ROI on the image. The block based PCA primarily applied on the ROI region whereas General PCA was applied to non-ROI region. From this work, it's found that region-based PCA performs much better than the PCA algorithm with regards to image quality, yielding similar compression ratio as the PCA algorithm.