ROI Based Image Compression of Medical Images
Amandeep Kaur, Monica Goyal · 2014
Medical imaging is one of the best techniques for monitoring the person’s health condition which is used widely nowadays. One of the problems that physicians encounter with it to store the medical images. This storage occupy more area for storing images long time as there is need to keep the record of numerous patients. So there is need to compress the image to be resolved in a variety of medical images, including radiography, magnetic resonance (MR), mammography, and ultrasound images, X-Rays, Brain MRI, CT images and so on. The rapid and reliable digital transmission and storage of medical and biomedical images would be tremendous boon to the practice of medicine. So this long term rapid transmission is prohibitive without image compression, to reduce the size of files. To make the Medical Images more useful and process able, there is need to reduce the transmission time and storage space for the images. The image may become more visual too, by compressed as it will also help to reduce transmission errors as less data will be transmit, also reduce the cost. Majority of the traditional methods lacks in terms of low compression ratio with lossless technique, some methods having problem of lost data when lossy compression technique is used. In vector quantization technique there is low PSNR values. Further, some methods provide high mean square errors. MAXSHIFT, SPIHT, and General scaling methods having low complexity but provide better results. DCT algorithm works efficiently to compress the image; it applied on the image blocks. If the blocks are too large then local features are no longer exploited, but if the blocks are too small then the images are not effectively correlated. So the size of the blocks is important as they determine the effectiveness of the transform over the image. But still there is not a single algorithm giving optimized values on above factors. Here our proposed work dealt with the investigation and implementation of traditional compression algorithm of Images of Medical domain, as a preliminary step and then proposed a compressed algorithm, known as a region growing algorithm to overcome the above adaptive problems being faced by traditional and also generated a quantitative analysis result of compression with various parameters PSNR and MSE. This proposed technique give best performance in terms of computation and speed of computation is high. Furthermore, resulting parameters (such as PSNR, MSE, and entropy) are calculated and compared with existing algorithm and observed that the proposed algorithm has better performance.