Research map / Detection and segmentation / CNN features
Geometry-Aware Distillation for Indoor Semantic Segmentation
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 |
|---|---|---|---|
| Use the proposed RGB-only geometry-aware distillation method instead of RDFNet | NYU-Dv2 semantic segmentation · NYU-Dv2 | mIoU: 50.1 -> 59.6 | Helped |
| Use the proposed RGB-only geometry-aware distillation method instead of RDFNet | SUN RGBD semantic segmentation · SUN RGBD | mIoU: 47.7 -> 54.5 | Helped |
| Add the proposed semantic loss (ablation) | NYU-Dv2 semantic-only model · NYU-Dv2 | mIoU: 36.0 -> 40.6 | Helped |
| Use learned depth-aware embeddings rather than HHA depth input (ablation) | NYU-Dv2 semantic segmentation · NYU-Dv2 | mIoU: 42.4 -> 44.7 | Helped |
| Add feature propagation from the depth branch (ablation) | NYU-Dv2 model with learned depth embeddings · NYU-Dv2 | mIoU: 44.7 -> 45.4 | Helped |
| Use geometry-aware propagation instead of vanilla convolution for fusion (ablation) | NYU-Dv2 model with depth feature propagation · NYU-Dv2 | mIoU: 55.1 -> 56.0 | Helped |
| Add skip pyramid fusion (ablation) | NYU-Dv2 model with geometry-aware propagation · NYU-Dv2 | mIoU: 56.0 -> 59.6 | Helped |
| Increase the amount of depth supervision used for training (ablation) | Proposed model · NYU-Dv2 | mIoU: 40.6 -> 41.1 -> 45.6 -> 51.6 -> 55.2 -> 59.6 | Helped |
Closest papers by meaning
- 3D-to-2D Distillation for Indoor Scene Parsing (CVPR 2021 · similarity 0.88)
- Geometric Exploitation for Indoor Panoramic Semantic Segmentation (NeurIPS 2024 · similarity 0.86)
- Geometry-Guided Semantic Reconstruction for 3D Instance Segmentation (NeurIPS 2026 · similarity 0.85)
- PDDM: Pseudo Depth Diffusion Model for RGB-PD Semantic Segmentation Based in Complex Indoor Scenes (AAAI 2025 · similarity 0.85)
- An End-to-End Robust Point Cloud Semantic Segmentation Network with Single-Step Conditional Diffusion Models (CVPR 2025 · similarity 0.84)
- Edge-Aware 3D Instance Segmentation Network with Intelligent Semantic Prior (CVPR 2024 · similarity 0.84)
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.