원문정보
초록
영어
The main drawback of the phase congruency feature employed in the feature similarity index (FSIM) image quality assessment (IQA) algorithm is its low computational efficiency. In this paper, a novel fast feature similarity index (FFSIM) for image quality assessment is proposed. Based on the fact that human visual system (HVS) responds to the brightness stimulus mainly complying with Weber's law, the proposed FFSIM only performs spatial filtering to quickly calculate the contrast between the current pixel and its background, which is used to compute Weber visual salience similarity and a weighting coefficient in pooling stage after applied nonlinear mapping. Weber contrast and the gradient magnitude play complementary roles in characterizing the image local quality. After obtaining the local quality map, we use Weber weighting coefficient again as a weighting coefficient to derive a single quality score. As such, the multi-scale version of the FFSIM algorithm, i.e., MS-FFSIM is also proposed, which complies with the spatial frequency response characteristics of the HVS system. Extensive experiments performed on six publicly available IQA databases demonstrate that the proposed FFSIM and MS-FFSIM can achieve higher consistency with the subjective evaluations than state-of-the-art IQA metrics and the computational efficiency is greatly improved as well.
목차
1. Introduction
2. FSIM
3. Fast FSIM
3.1. Basic Idea
3.2. Weber's Law
3.3. Weber Contrast Visual Saliency
3.4. FFSIM and its Multi-Scale Extension
4. Experiments and Results
4.1. Databases and Criteria for Comparison
4.2. Validation and Comparison
5. Conclusions
References