Research map / 3D vision / Depth and stereo
RGB-D Local Implicit Function for Depth Completion of Transparent Objects
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 the iterative depth refinement model (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.028 -> 0.025 | Helped |
| Use RGB and depth together rather than depth alone (ablation) | LIDF first stage · ClearGrasp Real-novel | RMSE: 0.052 -> 0.028 | Helped |
| Use RGB and depth together rather than RGB alone (ablation) | LIDF first stage · ClearGrasp Real-novel | RMSE: 0.062 -> 0.028 | Helped |
| Add ray direction and ray-voxel intersection information (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.050 -> 0.028 | Helped |
| Add positional encoding (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.032 -> 0.028 | Helped |
| Use 4^3 voxels rather than 8^3 (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.028 -> 0.030 | No clear effect |
| Use 16^3 voxels rather than 8^3 (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.028 -> 0.038 | Hurt |
| Use argmax pooling instead of weighted-sum pooling (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.032 -> 0.028 | Helped |
| Learn candidate-point offsets instead of sampling points (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.038 -> 0.028 | Helped |
| Train with ClearGrasp and Omniverse data instead of Omniverse data alone (ablation) | LIDF · ClearGrasp Real-novel | RMSE: 0.037 -> 0.028 | Helped |
| Replace ClearGrasp optimization with the proposed method | Transparent-object depth completion · ClearGrasp test set | inference time per image: 1.82 -> 0.09 second | Helped |
Closest papers by meaning
- Deep Depth Completion of a Single RGB-D Image (CVPR 2018 · similarity 0.90)
- From Transparent to Opaque: Rethinking Neural Implicit Surfaces with $\alpha$-NeuS (NeurIPS 2024 · similarity 0.87)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.86)
- Zero-shot Depth Completion via Test-time Alignment with Affine-invariant Depth Prior (AAAI 2025 · similarity 0.86)
- Differentiable Neural Surface Refinement for Modeling Transparent Objects (CVPR 2024 · similarity 0.85)
- RaySt3R: Predicting Novel Depth Maps for Zero-Shot Object Completion (NeurIPS 2025 · similarity 0.85)
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.