Lossy compression for medical images by using ideal cross-point regions and Lloyd’s algorithm

Abstract

This paper presents a lossy compression scheme for images by using the theory of cross-point regions and Lloyd’s algorithm with non-uniform quantization. The main idea of the method is that the application of the non-uniform quantization of Lloyd’s algorithm over Ideal Cross-point Regions (ICR). This quantization is very effective because the process of coding to make the code word uses the probability of data bits being optimized by the theory of cross-point regions. This means it can increase the compression ratio with better image quality—higher Peak Signal to Noise Ratio (PSNR) and less Mean Square Error (MSE). The algorithm can be implemented in storing patients’ medical imaging from Magnetic Resonant Imaging (MRI), Computed Tomography (CT) scanner, and for real time processing such as telemedicine. ICR can be also applied in connection with other quantization methods to improve compression scheme.