Full Reference Image Quality Assessment by Considering Intra-Block Structure and Inter-Block Texture.

Junfeng Yang,Wei Zhang,Xiaolong Li,Ting Zhou,Bo Ou

DOI: https://doi.org/10.1109/access.2020.3028282

IF: 3.9

2020-01-01

IEEE Access

Abstract:The purpose of the image quality assessment is to evaluate image quality consistently with human's subjective evaluation. In image quality assessment, many structural features have been used to measure the quality degradation of an image. However, most of the existing image quality assessment methods are block-based, which ignores the features among neighboring blocks. In this paper, we argue that the human visual system perceives distortions not only depends on local structural (intra-block structure) distortions, but also relates to the structural distortions of their neighborhoods (inter-block texture). Based on this insight, we propose a novel image quality assessment method, called the diffusion speed structure similarity (DSSIM), by considering both intra-block structure and inter-block texture. Specifically, to characterize the inter-block texture, we devise a novel visual feature based on the image diffusion speed. To measure the changes of the intra-block structure, we adopt the image gradient magnitude. Furthermore, to differentiate the importance of a local region, we devise a weighting function based on the image diffusion speed. Extensive experimental results on six benchmark databases demonstrate that our proposed method yields a better performance than the state-of-the-art methods.