Research map / Detection and segmentation / Weakly supervised detection and segmentation
Object Instance Mining for Weakly Supervised 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 instance reweighted loss to OIM (ablation) | OIM on VOC2007 test · VOC2007 test | mAP (%): 48.2 -> 50.1 | Helped |
| Combine spatial and appearance graphs (ablation) | OIM on VOC2007 test · VOC2007 test | mAP (%): 34.8 -> 48.2 | Helped |
| Use the appearance graph rather than the no-graph configuration (ablation) | OIM on VOC2007 test · VOC2007 test | mAP (%): 34.8 -> 46.7 | Helped |
| Use the spatial graph rather than the no-graph configuration (ablation) | OIM on VOC2007 test · VOC2007 test | mAP (%): 34.8 -> 42.2 | Helped |
| Use instance reweighted loss in OICR (ablation) | OICR on VOC2007 test · VOC2007 test | mAP (%): 41.2 -> 43.4 | Helped |
| Set alpha1=1 and alpha2=1 instead of alpha1=5 and alpha2=2 (ablation) | OIM+IR on VOC2007 test · VOC2007 test | mAP (%): 50.1 -> 43.4 | Hurt |
| Set alpha1=2 and alpha2=2 instead of alpha1=5 and alpha2=2 (ablation) | OIM+IR on VOC2007 test · VOC2007 test | mAP (%): 50.1 -> 49.3 | Hurt |
| Set alpha1=10 and alpha2=5 instead of alpha1=5 and alpha2=2 (ablation) | OIM+IR on VOC2007 test · VOC2007 test | mAP (%): 50.1 -> 48.4 | Hurt |
| Increase beta from 0.2 to 0.5 (ablation) | OIM+IR on VOC2007 test · VOC2007 test | mAP (%): 50.1 -> 48.0 | Hurt |
| Increase beta from 0.2 to 0.8 (ablation) | OIM+IR on VOC2007 test · VOC2007 test | mAP (%): 50.1 -> 46.3 | Hurt |
| Add OIM and instance reweighted loss instead of using C-MIL | WSOD on VOC2007 test · VOC2007 test | mAP (%): 50.5 -> 50.1 | Hurt |
| Use OIM+IR with Fast R-CNN instead of C-MIL+FRCNN | VOC2007 test · VOC2007 test | mAP (%): 53.1 -> 52.6 | Hurt |
Closest papers by meaning
- Instance Mining with Class Feature Banks for Weakly Supervised Object Detection (AAAI 2021 · similarity 0.93)
- Instance-Aware, Context-Focused, and Memory-Efficient Weakly Supervised Object Detection (CVPR 2020 · similarity 0.90)
- Multiple Instance Detection Network With Online Instance Classifier Refinement (CVPR 2017 · similarity 0.90)
- Weakly Supervised Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.90)
- Weakly Supervised Few-Shot Object Detection with DETR (AAAI 2024 · similarity 0.88)
- Relational Matching for Weakly Semi-Supervised Oriented Object Detection (CVPR 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.