Abstract
In real applications, obtained depth images are incomplete; therefore, depth image inpainting is studied here. A novel model that is characterised by both a low-rank structure and nonlocal self-similarity is proposed. As a double constraint, the low-rank structure and nonlocal self-similarity can fully exploit the features of single-depth images to complete the inpainting task. First, according to the characteristics of pixel values, we divide the image into blocks, and similar block groups and three-dimensional arrangements are then formed. Then, the variable splitting technique is applied to effectively divide the inpainting problem into the sub-problems of the low-rank constraint and nonlocal self-similarity constraint. Finally, different strategies are used to solve different sub-problems, resulting in greater reliability. Experiments show that the proposed algorithm attains state-of-the-art performance.
Highlights
With the rapid development of RGB-D sensors [1,2,3,4,5,6], such as the Kinect sensor, colour images and depth images can be obtained simultaneously
Depth images are of low quality and have black holes
Black holes represent missing depth information, and the black-hole filling problem is solved via depth image inpainting
Summary
With the rapid development of RGB-D (red green blue-depth) sensors [1,2,3,4,5,6], such as the Kinect sensor, colour images and depth images can be obtained simultaneously. Depth image inpainting methods can be divided into two categories according to whether the corresponding colour images are guided. Liu et al [11] proposed a robust optimisation framework for colour image-guided depth image restoration This method performs well in suppressing texture artefacts. Shen et al [14] proposed the inpainting method using a weighted joint bilateral filter and fast marching This method has obtained the best performance in improving depth images by producing smooth and edge regions. Lu et al [16] proposed a method of inpainting depth images through similar patches in a matrix and enforced low-rank subspace constraints, thereby attaining good performance. We use different strategies to solve depth image inpainting: weighted Schatten p-norm minimisation as the low-rank constraint and nonlocal statistical modelling as the nonlocal self-similarity constraint.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
