Share Email Print
cover

Proceedings Paper

The enhancement of depth estimation based on multi-scale convolution kernels
Author(s): Heng Hua; Xinzhu Sang; Xiyu Tian; Wanqi Sun; Duo Chen; Peng Wang
Format Member Price Non-Member Price
PDF $17.00 $21.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Depth prediction is essential for three-dimensional optical displays. The accuracy of the depth map influences the quality of virtual viewpoint synthesis. Due to the relatively simple end-to-end structures of CNNs, the performance for poor and repetitive texture is barely satisfactory. In consideration of the shortage of existing network structures, the two main structures are proposed to optimize the depth map. (i) Inspired by GoogLeNet, the inception module is added at the beginning of the network. (ii) Assuming that the disparity map has only horizontal disparity, two sizes of rectangular convolution kernels are introduced to the network structure. Experimental results demonstrate that our structures of the CNN reduce the error rate from 19.23% to 14.08%.

Paper Details

Date Published: 7 November 2018
PDF: 7 pages
Proc. SPIE 10817, Optoelectronic Imaging and Multimedia Technology V, 1081710 (7 November 2018); doi: 10.1117/12.2500775
Show Author Affiliations
Heng Hua, Beijing Univ. of Posts and Telecommunications (China)
Xinzhu Sang, Beijing Univ. of Posts and Telecommunications (China)
Xiyu Tian, Beijing Univ. of Posts and Telecommunications (China)
Wanqi Sun, Beijing Univ. of Posts and Telecommunications (China)
Duo Chen, Beijing Univ. of Posts and Telecommunications (China)
Peng Wang, Beijing Univ. of Posts and Telecommunications (China)


Published in SPIE Proceedings Vol. 10817:
Optoelectronic Imaging and Multimedia Technology V
Qionghai Dai; Tsutomu Shimura, Editor(s)

© SPIE. Terms of Use
Back to Top