Research map / Detection and segmentation / Open-set and few-shot detection
ZeroWaste Dataset: Towards Deformable Object Segmentation in Cluttered Scenes
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 |
|---|---|---|---|
| add TACO-cropped rare metal and rigid plastic objects to the training frames (ablation) | DeepLabV3+ · ZeroWaste-f test | mIoU: 52.13 -> 52.50 | Helped |
| add TACO-cropped rare metal and rigid plastic objects to the training frames (ablation) | DeepLabV3+ · ZeroWaste-f test | Pixel accuracy: 91.38 -> 91.44 | Helped |
| add unlabeled ZeroWaste-s examples during training (ablation) | ReCo · ZeroWaste-f test | mIoU: 52.28 -> 44.12 | Hurt |
| add unlabeled ZeroWaste-s examples during training (ablation) | CCT · ZeroWaste-f test | mIoU: 29.32 -> 32.49 | Helped |
| use PuzzleCAM rather than CAM for weakly supervised segmentation | ZeroWaste-w weakly supervised segmentation · ZeroWaste-f test | mIoU: 19.1 -> 28.46 | Helped |
| use EPS rather than PuzzleCAM for weakly supervised segmentation | ZeroWaste-w weakly supervised segmentation · ZeroWaste-f test | mIoU: 28.46 -> 34.06 | Helped |
| pretrain Mask R-CNN on TACO rather than COCO | Mask R-CNN · ZeroWaste-f test | AP: 21.97 -> 24.32 | Helped |
| train with before/after image-level labels rather than class-wise image-level labels | EPS · ZeroWaste-f test | mIoU: 13.91 -> 34.06 | Helped |
| use ReCo instead of DeepLabV3+ | fully supervised semantic segmentation · ZeroWaste-f test | mIoU: 52.13 -> 52.28 | Mixed |
Closest papers by meaning
- UrbanWaste: In-the-Bin Dataset for Waste Disposal Inspection with Multi-Granularity Hierarchical Labels (AAAI 2025 · similarity 0.89)
- WISNet: Pseudo Label Generation on Unbalanced and Patch Annotated Waste Images (CVPR 2025 · similarity 0.86)
- Infusing Synthetic Data with Real-World Patterns for Zero-Shot Material State Segmentation (NeurIPS 2024 · similarity 0.84)
- DiffCut: Catalyzing Zero-Shot Semantic Segmentation with Diffusion Features and Recursive Normalized Cut (NeurIPS 2024 · similarity 0.81)
- Locate Anything on Earth: Advancing Open-Vocabulary Object Detection for Remote Sensing Community (AAAI 2025 · similarity 0.80)
- unMORE: Unsupervised Multi-Object Segmentation via Center-Boundary Reasoning (ICML 2025 · similarity 0.79)
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.