Research map / 3D vision / Depth and stereo
Indoor Obstacle Discovery on Reflective Ground via Monocular Camera
Feng Xue, Yicong Chang, Tianxi Wang, Yu Chen Zhou, Anlong Ming
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 geometry features (AGR) to appearance features (ablation) | AR · Reflective-ground test | ITPR: 64.47 -> 77.46 | Improved |
| add appearance features to geometry features (ablation) | AGR · Reflective-ground test | TPR: 67.67 -> 74.23 | Improved |
| add homography error to geometry features (ablation) | AGR w/o HE · Reflective-ground test | ITPR: 76.46 -> 76.83 | Improved |
| add deviation angle to geometry features (ablation) | AGR w/o DA · Reflective-ground test | ITPR: 73.24 -> 76.83 | Improved |
| add proposed features and probability map instead of TOD | TOD · Reflective-ground test | ITPR: 61.72 -> 77.46 | Improved |
| add proposed features instead of TOD | TOD · Reflective-ground test | MIFP: 3.46 -> 1.87 | Improved |
| use Top 50 candidates instead of Top 10 in probability map (ablation) | AGR+AR · Reflective-ground test | ITPR: 71.47 -> 77.46 | Improved |
| use Top 100 instead of Top 50 in probability map (ablation) | AGR+AR · Reflective-ground test | ITPR: 77.46 -> 78.37 | No clear effect |
| use Top 500 instead of Top 50 in probability map (ablation) | AGR+AR · Reflective-ground test | ITPR: 77.46 -> 76.55 | Hurt |
| increase motion blur kernel from 0 to 19 (ablation) | AGR+AR · Reflective-ground test | ITPR: 77.46 -> 57.32 | Hurt |
| add translation noise of 5-10% to robot motion (ablation) | AGR+AR · Reflective-ground test | ITPR: 77.46 -> 77.01 | No clear effect |
| add both translation and rotation noise at maximum magnitude (ablation) | AGR+AR · Reflective-ground test | ITPR: 77.46 -> 76.64 | Hurt |
Closest papers by meaning
- Footprints and Free Space From a Single Color Image (CVPR 2020 · similarity 0.84)
- Monocular 3D Object Detection with Depth from Motion (ECCV 2022 · similarity 0.82)
- MonoGround: Detecting Monocular 3D Objects From the Ground (CVPR 2022 · similarity 0.82)
- Can NeRFs "See" without Cameras? (NeurIPS 2025 · similarity 0.80)
- Layered Motion Fusion: Lifting Motion Segmentation to 3D in Egocentric Videos (CVPR 2025 · similarity 0.80)
- SeeGround: See and Ground for Zero-Shot Open-Vocabulary 3D Visual Grounding (CVPR 2025 · similarity 0.80)
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.
