Research map / Detection and segmentation / CNN features
Joint Task-Recursive Learning for Semantic Segmentation and Depth Estimation
Zhenyu Zhang, Zhen Cui, Chunyan Xu, Zequn Jie, Xiang Li, Jian Yang
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 learning of segmentation with depth (ablation) | TRL (ResNet-18) · NYU Depth v2 segmentation | mean IoU: 42.0 -> 45.0 | Improved |
| add joint learning of segmentation with depth (ablation) | TRL (ResNet-18) · SUN-RGBD segmentation | mean IoU: 43.5 -> 46.3 | Improved |
| add joint learning of depth with segmentation (ablation) | TRL (ResNet-18) · NYU Depth v2 depth | rms: 0.547 -> 0.510 | Improved |
| add joint learning of depth with segmentation (ablation) | TRL (ResNet-18) · SUN-RGBD depth | rms: 0.517 -> 0.468 | Improved |
| add task-attentional modules (TAM) (ablation) | TRL (ResNet-18) · NYU Depth v2 depth | rms: 0.526 -> 0.510 | Improved |
| add task-attentional modules (TAM) (ablation) | TRL (ResNet-18) · NYU Depth v2 segmentation | mean IoU: 43.6 -> 45.0 | Improved |
| add historical experience propagation (ablation) | TRL (ResNet-18) · NYU Depth v2 depth | rms: 0.540 -> 0.510 | Improved |
| add historical experience propagation (ablation) | TRL (ResNet-18) · NYU Depth v2 segmentation | mean IoU: 42.2 -> 45.0 | Improved |
| add gate unit in TAM (ablation) | TRL (ResNet-18) · NYU Depth v2 segmentation | mean IoU: 44.7 -> 45.0 | Improved |
| add gate unit in TAM (ablation) | TRL (ResNet-18) · NYU Depth v2 depth | rms: 0.515 -> 0.510 | Improved |
| use full TRL instead of baseline with parallel decoders and cross-stitch (ablation) | TRL (ResNet-18) · NYU Depth v2 depth | rms: 0.545 -> 0.510 | Improved |
| predict at finer scale (scale-1 to final) (ablation) | TRL (ResNet-18) · NYU Depth v2 segmentation | mean IoU: 40.3 -> 45.0 | Improved |
Closest papers by meaning
- Joint Task-Recursive Learning for RGB-D Scene Understanding (TPAMI 2019 · similarity 0.93)
- Learning To Relate Depth and Semantics for Unsupervised Domain Adaptation (CVPR 2021 · similarity 0.90)
- Recurrent Scene Parsing With Perspective Understanding in the Loop (CVPR 2018 · similarity 0.89)
- SDC-Depth: Semantic Divide-and-Conquer Network for Monocular Depth Estimation (CVPR 2020 · similarity 0.88)
- Look Deeper into Depth: Monocular Depth Estimation with Semantic Booster and Attention-Driven Loss (ECCV 2018 · similarity 0.88)
- Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling (CVPR 2024 · 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.
