A comprehensive intelligent compression method on DICOM images

Abstract

Virtual medical imaging has overgrown in recent years and later implemented in such situations - all of the radiological modalities which include Computed tomography (CT scanners), Magnetic resonance imaging (MRI), ultrasound (US), positron emission tomography (PET), X-Ray (radiographs) made employing more than one providers and hosted by one or many websites which communicated over the DICOM network. The DICOM images required huge hard disk space and excellent transfer speed, which need to compress the DICOM images for effective capacity and transmission over the internet. The compression process is applied through utilized a recurrent neural network algorithm establishing Trainscg as the activation function. Extraordinary high-quality metrics like mean squared error (MSE), Peak signal-to-noise ratio (PSNR), Compression ratio (CR), and Compression time are computed on several medical test images. The proposed compression method shows better experimental results than the existing techniques based on performance parameters except for the compression time for the large image only.