Name
Abstract | ||
|---|---|---|
Human Visual System (HVS) is the terminal receiver of digital images, and the perception of image quality is based on human visual characteristics. As is well known, HVS is highly adapted to extract structural information from the scene. However, existing image quality assessment (IQA) methods, which aim to measure the image quality consistently with human perception, have not well exploited the visual structural saliency. Here, a novel method is proposed, which improves the present situation by introducing the structural saliency model (SSM). The SSM is implemented by the global probability of boundary map which provides a hierarchical structural information. The hierarchical structural information truly reflects the discriminative response of HVS to the different image structural stimuli. Meanwhile, we also adopt the phase congruency (PC) and the gradient magnitude (GM) information. The former can accurately characterize the significance of image features, and the latter is also a useful primary feature of image. They are two commonly used sub-indexes and have been verified effective in many other IQA researches. Extensive experiments performed on three publicly available image databases demonstrate that the structural saliency model is accurate in assigning visual importance, and a comprehensive performance improvement is witnessed. |
Year | DOI | Venue |
|---|---|---|
2014 | 10.1109/ICDSP.2014.6900714 | DSP |
Keywords | DocType | ISSN |
image processing,human visual system,structural saliency model,digital images,visual databases,human visual characteristics,terminal receiver,visual structural saliency,image features,gradient magnitude information,structural information,comprehensive performance,image quality perception,gradient magnitude,image quality assessment,IQA methods,hierarchical structural information,structural saliency,phase congruency | Conference | 1546-1874 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ziran Zhang | 1 | 0 | 0.34 |
| Jianhua Zhang | 2 | 25 | 5.97 |
| Xiaoyan Wang | 3 | 0 | 1.35 |
| Qiu Guan | 4 | 43 | 9.92 |
| Sheng-Yong Chen | 5 | 1077 | 114.06 |