An Efficient and Accurate Stereo Matching Algorithm Based on Convolutional Neural Network

被引：0
|
作者：
Zhang W. [1 ]
Shao X. [1 ]
Yang W. [1 ]
Guo M. [1 ]
Jing N. [1 ]
机构：
[1] School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing
来源：
Jisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/Journal of Computer-Aided Design and Computer Graphics | 2020年 / 32卷 / 01期
关键词：
Attention mechanism; Convolutional neural network; Disparity map; Stereo matching;
D O I：
10.3724/SP.J.1089.2020.17823
中图分类号：
学科分类号：
摘要：
Aiming at the problems of huge parameters and inaccurate accuracy in stereo matching algorithm based on convolution neural network (CNN), this paper proposes an efficient and accurate stereo matching algorithm based on CNN. Firstly, a feature extraction network combining multi-dimensional context information is designed to improve the matching accuracy of ill-posed regions. Secondly, the existing similarity calculation steps are improved to reduce the amount of the network parameter while ensuring the matching accuracy. Finally, a lightweight attention-based disparity refinement algorithm is proposed, which focuses on and modifies the erroneous pixels of the initial disparity map from the channel and spatial dimensions. Compared with GC-Net on the standard dataset Sceneflow, the proposed algorithm improves the matching accuracy by more than 50% while reducing 14% parameters. © 2020, Beijing China Science Journal Publishing Co. Ltd. All right reserved.
引用
收藏
页码：45 / 53
页数：8
相关论文
共 20 条
  • [1] Scharstein D., Szeliski R., A taxonomy and evaluation of dense two-frame stereo correspondence algorithms, International Journal of Computer Vision, 47, 1-3, pp. 7-42, (2002)
  • [2] Zbontar J., LeCun Y., Computing the stereo matching cost with a convolutional neural network, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1592-1599, (2015)
  • [3] Luo W.J., Schwing A.G., Urtasun R., Efficient deep learning for stereo matching, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5695-5703, (2016)
  • [4] Guney F., Geiger A., Displets: resolving stereo ambiguities using object knowledge, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4165-4175, (2015)
  • [5] Shaked A., Wolf L., Improved stereo matching with constant highway networks and reflective confidence learning, Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 6901-6910, (2017)
  • [6] Mayer N., Ilg E., Hausser P., Et al., A large dataset to train convolutional networks for disparity, optical flow, and scene flow estimation, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4040-4048, (2016)
  • [7] Pang J.H., Sun W.X., Ren J.S., Et al., Cascade residual learning: a two-stage convolutional neural network for stereo matching, Proceedings of the IEEE International Conference on Computer Vision Workshops, pp. 887-895, (2017)
  • [8] Kendall A., Martirosyan H., Dasgupta S., Et al., End-to-end learning of geometry and context for deep stereo regression, Proceedings of the IEEE International Conference on Computer Vision, pp. 66-75, (2017)
  • [9] Chang J.R., Chen Y.S., Pyramid stereo matching network, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5410-5418, (2018)
  • [10] Smolyanskiy N., Kamenev A., Birchfield S., On the importance of stereo for accurate depth estimation: an efficient semi-supervised deep neural network approach, Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, pp. 1120-1128, (2018)
12