Research map / 3D vision / Depth and stereo
Occlusion-Aware Cost Constructor for Light Field Depth Estimation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add pixel modulation to weight view contributions (ablation) | OACC-Net · 4D LF benchmark average | BadPix0.07: 4.494 -> 2.981 | Helped |
| add pixel modulation to weight view contributions (ablation) | OACC-Net · 4D LF benchmark average | MSE (×100): 1.572 -> 1.236 | Helped |
| generate an occlusion mask from the initial estimated disparity and update it during inference (ablation) | OACC-Net · 4D LF benchmark average | BadPix0.07: 3.143 -> 2.981 | Helped |
| generate an occlusion mask from the initial estimated disparity and update it during inference (ablation) | OACC-Net · 4D LF benchmark average | MSE (×100): 1.293 -> 1.236 | Helped |
| increase inference iterations from two to three (ablation) | OACC-Net · 4D LF benchmark average | BadPix0.07: 2.981 -> 2.970 | No clear effect |
| increase inference iterations from two to three (ablation) | OACC-Net · 4D LF benchmark average | MSE (×100): 1.236 -> 1.240 | No clear effect |
| increase inference iterations from two to four (ablation) | OACC-Net · 4D LF benchmark average | BadPix0.07: 2.981 -> 2.975 | No clear effect |
| increase inference iterations from two to four (ablation) | OACC-Net · 4D LF benchmark average | MSE (×100): 1.236 -> 1.237 | No clear effect |
| use occlusion masks generated from ground-truth disparities (ablation) | OACC-Net · 4D LF benchmark average | MSE (×100): 1.236 -> 1.088 | Helped |
| replace shift-and-concatenate cost construction with OACC convolutions | OACC-Net · 9×9 LF, 512×512 input | cost construction time (s): 2.741 -> 0.004 | Helped |
| increase the occlusion-mask decay rate (ablation) | OACC-Net · scenes with heavy occlusions and large noise | MSE | Mixed |
Closest papers by meaning
- Attention-Based View Selection Networks for Light-Field Disparity Estimation (AAAI 2020 · similarity 0.87)
- Trust Your Model: Light Field Depth Estimation With Inline Occlusion Handling (CVPR 2018 · similarity 0.86)
- Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation (NeurIPS 2020 · similarity 0.86)
- Learning Fused Pixel and Feature-Based View Reconstructions for Light Fields (CVPR 2020 · similarity 0.86)
- Robust Light Field Depth Estimation for Noisy Scene With Occlusion (CVPR 2016 · similarity 0.86)
- Convolutional Networks for Shape From Light Field (CVPR 2016 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.