A New Wavelet Visual Weighting Model to Optimize Image Coding and Quality Evaluation Based on the Human Psychovisual Quality Properties

In this work, we suggest the development of two new image coders, namely EVIC (Embedded Visual-based image coder) based on the perceptual model and a designed weighting model and its optimized version VOEVIC (Visual Optimized Embedded Visual-based image coder) in order to obtain an improved perceptual quality compared to the standard SPIHT coder, for a given bit rate. Among other things, we propose a new VWVDP quality evaluator and its optimized version VOWVDP, and suggest the validation of objective metrics for evaluating coding quality using Human Visual System (HVS) quality criteria. These metrics incorporate visual weighting to eliminate any undesirable components invisible to the human visual cortex and retain only those that are perceptible in the cortical domain, and finally evaluate the differences by offering a map of visible differences VDM and a quality factor PS (Probability Score) for visual coding evaluation. This paper sets the objective of optimizing PAYLAOD in IoT platforms deploying computational intelligence for object detection and recognition with a view to offering green solutions to smart city applications, in areas such as health, transport, agriculture, defense and human security, for PAYLOAD size reduction while optimizing quality and reducing the cost of energy consumption and multimedia data transport times.