Research map / Detection and segmentation / Weakly supervised detection and segmentation
Pseudo Mask Augmented Object Detection
Read the paperCheck 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 coarse pseudo-mask segmentation supervision (ablation) | Faster R-CNN with VGG-16 · PASCAL VOC 2007 test | mAP: 74.5 -> 74.9 | Helped |
| Add iterative graph-cut pseudo-mask refinement (ablation) | PAD without detection feedback, VGG-16 · PASCAL VOC 2007 test | mAP: 74.9 -> 75.7 | Helped |
| Add global segmentation feedback (ablation) | PAD with refined masks, VGG-16 · PASCAL VOC 2007 test | mAP: 75.7 -> 76.0 | Helped |
| Add instance-level segmentation feedback (ablation) | PAD with refined masks, VGG-16 · PASCAL VOC 2007 test | mAP: 75.7 -> 76.2 | Helped |
| Add both global and instance segmentation feedback (ablation) | PAD with refined masks, VGG-16 · PASCAL VOC 2007 test | mAP: 75.7 -> 77.0 | Helped |
| Increase iterative pseudo-mask refinement from one to two to three iterations (ablation) | PAD, VGG-16 · PASCAL VOC 2007 test | mAP: 75.9 -> 76.4 -> 77.0 | Helped |
| Replace iterative graph-cut refinement with GrabCut masks (ablation) | PAD · PASCAL VOC 2012 SDS validation | mAP^r: 48.3 vs 58.5 | Hurt |
| Remove the 1D box loss (ablation) | PAD · PASCAL VOC 2012 SDS validation | mAP^r: 49.1 vs 58.5 | Hurt |
| Iteratively refine pseudo masks (ablation) | PAD · PASCAL VOC 2012 SDS validation | mAP^r: 44.3 -> 52.1 -> 58.0 -> 58.5 | Helped |
| Use bounding-box-supervised pseudo masks instead of ground-truth masks (ablation) | PAD · PASCAL VOC 2012 SDS validation | mAP^r: 64.5 vs 58.5 | Hurt |
Closest papers by meaning
- BoxTeacher: Exploring High-Quality Pseudo Labels for Weakly Supervised Instance Segmentation (CVPR 2023 · similarity 0.89)
- W2F: A Weakly-Supervised to Fully-Supervised Framework for Object Detection (CVPR 2018 · similarity 0.89)
- MaskLab: Instance Segmentation by Refining Object Detection With Semantic and Direction Features (CVPR 2018 · similarity 0.88)
- Mask DINO: Towards a Unified Transformer-Based Framework for Object Detection and Segmentation (CVPR 2023 · similarity 0.88)
- DI-MaskDINO: A Joint Object Detection and Instance Segmentation Model (NeurIPS 2024 · similarity 0.86)
- The Devil is in the Object Boundary: Towards Annotation-free Instance Segmentation using Foundation Models (ICLR 2024 · similarity 0.86)
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.