Laser & Optoelectronics Progress, Volume. 60, Issue 12, 1211002(2023)
Depth Estimation for Phase-Coding Light Field Based on Neural Network
Fig. 1. Three-branch light field depth estimation network
Fig. 2. Multi-scale feature extraction module
Fig. 3. Data augmentation by rotation
Fig. 4. Convergence curve of network training
Fig. 5. Comparison of the data results of each method on the test set. (a) MAE; (b) BP7; (c) BP5; (d) BP3
Fig. 6. Comparison of depth map results of various methods. (a) CAE; (b) OCC; (c) SPO; (d)REFOCUS; (e) EPINet; (f) proposed method; (g) ground-truth
Fig. 7. Comparison of error maps of the proposed method and EPINet
Fig. 8. Comparison of depth map results of the proposed network with/without center view
Fig. 9. Comparison of error map results of the proposed network with/without center view
|
|
|
|
Get Citation
Copy Citation Text
Chengzhuo Yang, Sen Xiang, Huiping Deng, Jing Wu. Depth Estimation for Phase-Coding Light Field Based on Neural Network[J]. Laser & Optoelectronics Progress, 2023, 60(12): 1211002
Category: Imaging Systems
Received: Mar. 29, 2022
Accepted: Jun. 14, 2022
Published Online: Jun. 5, 2023
The Author Email: Xiang Sen (xiangsen@wust.edu.cn)