Research map / 3D vision / Depth and stereo
Lens Parameter Estimation for Realistic Depth of Field Modeling
Dominique Piché-Meunier, Yannick Hold-Geoffroy, Jianming Zhang, Jean-François Lalonde
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 weight network to the least-squares lens parameter fit (ablation) | lens parameter estimation · BokehMe (DIML), out-of-domain | blur factor median error: 3.62 -> 3.49 | Improved |
| add weight network to the least-squares fit (ablation) | lens parameter estimation · SynthWorld and BokehMe | median error of lens parameters: 4.59 -> 4.55 (SynthWorld blur factor) | No clear effect |
| finetune disparity network on defocused SynthWorld and BokehMe data (ablation) | disparity network · BokehMe disparity | RMSE: 0.140 -> 0.043 | Improved |
| finetune disparity network on defocused data (ablation) | disparity network · SynthWorld disparity | RMSE: 0.152 -> 0.087 | Improved |
| replace weight network with direct regression of parameters (ablation) | lens parameter estimation · BokehMe (DIML) | blur factor median error: 3.49 -> 11.5 | Hurt |
| use RANSAC instead of weight network (ablation) | lens parameter estimation · BokehMe (DIML) | blur factor median error: 3.49 -> 4.60 | Hurt |
| use image gradients as weights instead of learned weights (ablation) | lens parameter estimation · BokehMe (DIML) | blur factor median error: 3.49 -> 3.62 | Hurt |
| use lens parameter fit on top of pixel maps from prior defocus methods | DMENet, DBENet, DID-ANet baselines · SynthWorld | blur factor median error: 17.12 vs 4.55 for ours | Hurt |
| reconstruct defocus from lens parameters and disparity instead of raw defocus output | defocus estimation · BokehMe defocus | RMSE (px): 2.22 -> 2.13 (BokehMe); 5.83 -> 6.36 (SynthWorld) | Mixed |
Closest papers by meaning
- Deep Optics for Monocular Depth Estimation and 3D Object Detection (ICCV 2019 · similarity 0.89)
- Deep Depth from Focal Stack with Defocus Model for Camera-Setting Invariance (IJCV 2023 · similarity 0.89)
- Focus on Defocus: Bridging the Synthetic to Real Domain Gap for Depth Estimation (CVPR 2020 · similarity 0.88)
- Shape Estimation from Shading, Defocus, and Correspondence Using Light-Field Angular Coherence (TPAMI 2016 · similarity 0.88)
- Multi-Focus Image Fusion via Explicit Defocus Blur Modelling (AAAI 2025 · similarity 0.87)
- Blurry-Edges: Photon-Limited Depth Estimation from Defocused Boundaries (CVPR 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.
