Name
Title | ||
|---|---|---|
Blind stereoscopic image quality assessment using 3D saliency selected binocular perception and 3D convolutional neural network |
Abstract | ||
|---|---|---|
The purpose of stereoscopic image quality assessment (SIQA) is to design an objective evaluation algorithm to automatically evaluate the quality of stereoscopic image. In this paper, we propose a blind SIQA method via 3D saliency selected binocular perception and 3D convolutional neural network (CNN). Given a pair of stereoscopic images, we first generate 3D saliency map by weighted average of 2D saliency map and depth saliency map. Then, when the value of 3D saliency map patches is higher than the setting threshold, these patches from left and right images are selected to feed to 3D-CNN to predict the perceived quality. Finally, the score of the distorted stereoscopic image is computed by the weighted average of the quality scores of these saliency image patches. Experimental results on LIVE 3D Phase I and Phase II databases show that our proposed method is robust and competitive with the state-of-the-art NR SIQA methods. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1007/s11042-022-12707-4 | Multimedia Tools and Applications |
Keywords | DocType | Volume |
Blind stereoscopic image quality assessment, Convolutional neural network, 3D saliency map, Summation and difference image, Cyclopean image | Journal | 81 |
Issue | ISSN | Citations |
13 | 1380-7501 | 0 |
PageRank | References | Authors |
0.34 | 34 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Chaofeng Li | 1 | 0 | 0.34 |
| LiXia Yun | 2 | 0 | 0.34 |
| Shoukun Xu | 3 | 1 | 1.71 |