Research map / 3D vision / Depth and stereo
Unified Depth Prediction and Intrinsic Image Decomposition from a Single Image via Joint Convolutional Neural Fields
Seungryong Kim, Kihong Park, Kwanghoon Sohn, Stephen Lin
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 joint network learning (shared activations between depth and intrinsic networks) (ablation) | JCNF · MPI Sintel depth | rel: 0.292 -> 0.183 | Improved |
| add the gradient scale network (ablation) | JCNF · MPI Sintel depth | rel: 0.271 -> 0.183 | Improved |
| add coarse-to-fine joint prediction (ablation) | JCNF · MPI Sintel depth | rel: 0.252 -> 0.183 | Improved |
| add joint network learning (shared activations between depth and intrinsic networks) (ablation) | JCNF · MPI Sintel intrinsic decomposition | MSE (avg.): 0.016 -> 0.007 | Improved |
| add the gradient scale network (ablation) | JCNF · MPI Sintel intrinsic decomposition | MSE (avg.): 0.011 -> 0.007 | Improved |
| add coarse-to-fine joint prediction (ablation) | JCNF · MPI Sintel intrinsic decomposition | MSE (avg.): 0.010 -> 0.007 | Improved |
| add joint network learning (shared activations between depth and intrinsic networks) (ablation) | JCNF · NYU v2 depth | rel: 0.216 -> 0.201 | Improved |
| add the gradient scale network (ablation) | JCNF · NYU v2 depth | rel: 0.210 -> 0.201 | Improved |
| add coarse-to-fine joint prediction (ablation) | JCNF · NYU v2 depth | rel: 0.208 -> 0.201 | Improved |
| pre-train on MPI Sintel or NYU v2 rather than train on Make3D itself | JCNF · Make3D depth (C1, depth < 70 m) | rel: 0.262 -> 0.273 | Hurt |
Closest papers by meaning
- Learning Depth from Single Monocular Images Using Deep Convolutional Neural Fields (TPAMI 2016 · similarity 0.92)
- Structured Attention Guided Convolutional Neural Fields for Monocular Depth Estimation (CVPR 2018 · similarity 0.89)
- Monocular Depth Estimation Using Neural Regression Forest (CVPR 2016 · similarity 0.88)
- Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue (ECCV 2016 · similarity 0.88)
- Monocular Depth Estimation Using Relative Depth Maps (CVPR 2019 · similarity 0.88)
- Self-Supervised Monocular Depth Estimation by Direction-aware Cumulative Convolution Network (ICCV 2023 · similarity 0.88)
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.
