基于三维视角的图像质量评价综述OA

Image quality assessment(IQA),as one of the core research directions in the field of image processing and computer vision,aims to achieve the automatic assessment of the visual quality of images and videos.The results directly affect the reliability and effectiveness of downstream applications,such as medical diagnosis,intelligent monitoring,remote sensing analysis,and multimedia content distribution.Most of the existing reviews focus on specific application scenarios or single technical paths.Although they provide a systematic review of local directions,they are still insufficient in comprehensively integrating different paradigms and revealing the evolution laws of cross-stage methods.To fill this gap,this paper proposes a structured classification framework with"three-dimensional perspective"perception,feature representation,prediction and mapping as the main line.This framework systematically summarizes and compares the existing IQA methods from three dimensions:the human visual perception mechanism,the feature information representa-tion method,and the mapping relationship of quality prediction.Based on this,this paper respectively sorts out the key issues and typical methods in each dimension,summarizes the main research results,and deeply analyzes the applicable scenarios,advantages and limitations of different methods.Finally,the key issues that urgently need to be addressed in the current IQA research are summarized,and the future development trends are prospected.