Research On Image Quality Assessment Based On Structure Similarity

Researches on image quality assessment are meaningful and basic for the image processing engineering. Considering the low computation complexity of traditional image quality measures, they are widely used. But they can not keep coherence with subject visual quality. The structure similarity index(SSIM) which is proposed recently, can reflect human visual characteristic better. But it shows some defects when evaluating the quality of blur images and noise images, the results don't match the HVS system.In this paper, a new image quality assessment algorithm based on dual-scale edge structure similarity (MDESSIM) is proposed. The edge information is appreciated sufficiently during the structure information in this algorithm. The distortion metric of edge structure is included in the assessment of the image quality. The history and current state of development and research of the image quality assessment are also introduced. And the characteristics of Human Visual System are discussed. The last experiment results demonstrate the DESSIM algorithm can keep better consistency with subjective perception for the wider kinds of images.