CAttNet: A Compound Attention Network for Depth Estimation of Light Field Images
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Depth estimation is one of the most complicated and difficult problems to deal with in the light field. In thispaper, a compound attention convolutional neural network (CAttNet) is proposed to extract depth maps fromlight field images. To make more effective use of the sub-aperture images (SAIs) of light field and reduce theredundancy in SAIs, we use a compound attention mechanism to weigh the channel and space of the featuremap after extracting the primary features, so it can more efficiently select the required view and the importantarea within the view. We modified various layers of feature extraction to make it more efficient and useful toextract features without adding parameters. By exploring the characteristics of light field, we increased thenetwork depth and optimized the network structure to reduce the adverse impact of this change. CAttNet canefficiently utilize different SAIs correlations and features to generate a high-quality light field depth map. Theexperimental results show that CAttNet has advantages in both accuracy and time.
Abstract
Depth estimation is one of the most complicated and difficult problems to deal with in the light field. In thispaper, a compound attention convolutional neural network (CAttNet) is proposed to extract depth maps fromlight field images. To make more effective use of the sub-aperture images (SAIs) of light field and reduce theredundancy in SAIs, we use a compound attention mechanism to weigh the channel and space of the featuremap after extracting the primary features, so it can more efficiently select the required view and the importantarea within the view. We modified various layers of feature extraction to make it more efficient and useful toextract features without adding parameters. By exploring the characteristics of light field, we increased thenetwork depth and optimized the network structure to reduce the adverse impact of this change. CAttNet canefficiently utilize different SAIs correlations and features to generate a high-quality light field depth map. Theexperimental results show that CAttNet has advantages in both accuracy and time.
참고문헌 (24)
- A. C. Tsai, "Occlusion resistant face detection and recognition system", Proceedings of 2020 8th International Conference on Orange Technology (ICOT), pp.1~4 (2020)
- J. Liu, "Survey of the image recognition based on deep learning network for autonomous driving car", Proceedings of 2020 5th International Conference on Information Science, Computer Technology and Transportation (ISCTT), pp.1~6 (2020)
- Xian-Feng Han, "Image-Based 3D Object Reconstruction: State-of-the-Art and Trends in the Deep Learning Era", IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(5), pp.1578~1604 (2021), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Hsiao-Chien Yang, "FADE: Feature Aggregation for Depth Estimation With Multi-View Stereo", IEEE Transactions on Image Processing, 29, pp.6590~6600 (2020), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Yongbing Zhang, "Light-Field Depth Estimation via Epipolar Plane Image Analysis and Locally Linear Embedding", IEEE Transactions on Circuits and Systems for Video Technology, 27(4), pp.739~747 (2017), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- A. Ak, "Investigating epipolar plane image representations for objective quality evaluation of light field images", Proceedings of 2019 8th European Workshop on Visual Information Processing (EUVIP), pp.135~139 (2019)
- W. Zhou, "Learning depth cues from focal stack for light field depth estimation", Proceedings of 2019 IEEE International Conference on Image Processing (ICIP), pp.1074~1078 (2019)
- C. Shin, "EpiNet: a fully-convolutional neural network using epipolar geometry for depth from light field images", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.4748~4757 (2018)
- K. Honauer, "Computer Vision–ACCV 2016", pp.19~34 (2017), Springer
- S. Woo, "Computer Vision–ECCV 2018", pp.3~19 (2018), Springer
- K. He, "Deep residual learning for image recognition", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.770~778 (2016)
- Z. Yu, "Line assisted light field triangulation and stereo matching", Proceedings of the IEEE International Conference on Computer Vision, pp.2792~2799 (2013)
- S. Heber, "Computer Vision–ECCV 2014", pp.751~767 (2014), Springer
- Jie Chen, "Accurate Light Field Depth Estimation With Superpixel Regularization Over Partially Occluded Regions", IEEE Transactions on Image Processing, 27(10), pp.4889~4900 (2018), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- M. W. Tao, "Depth from combining defocus and correspondence using light-field cameras", Proceedings of the IEEE International Conference on Computer Vision, pp.673~680 (2013)
- Sven Wanner, "Variational Light Field Analysis for Disparity Estimation and Super-Resolution", IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(3), pp.606~619 (2014), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Hao Sheng, "Occlusion-aware depth estimation for light field using multi-orientation EPIs", Pattern Recognition, 74, pp.587~599 (2018), Elsevier BV[DOI]
- J. Li, "EPI-neighborhood distribution based light field depth estimation", Proceedings of 2020IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp.2003~2007 (2020)
- Y. J. Tsai, "Attention-based view selection networks for lightfield disparity estimation", Proceedings of the AAAI Conference on Artificial Intelligence, 34(07), pp.12095~12103 (2020)
- Y. Li, "MANet: multi-scale aggregated network for light field depth estimation", Proceedings of 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp.1998~2002 (2020)
- Yan Li, "A Lightweight Depth Estimation Network for Wide-Baseline Light Fields", IEEE Transactions on Image Processing, 30, pp.2288~2300 (2021), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- G. Klambauer, "Self-normalizing neural networks", Advances in Neural Information Processing Systems, 30, pp.971~980 (2017)
- H. Schilling, "Trust your model : light field depth estimation with inline occlusion handling", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.4530~4538 (2018)
- Y. Luo, "Neural Information Processing", pp.642~652 (2017), Springer