Research map / Detection and segmentation / CNN features
Sequential Context Encoding for Duplicate Removal
Lu Qi, Shu Liu, Jianping Shi, Jiaya Jia
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 |
|---|---|---|---|
| Replace NMS with two-stage sequential context encoding | FPN · COCO val | mAP: 37.1 -> 38.3 | Improved |
| Replace NMS with two-stage sequential context encoding | Mask R-CNN · COCO val | mAP: 38.9 -> 40.2 | Improved |
| Replace NMS with two-stage sequential context encoding | PANet with DCN · COCO val | mAP: 43.7 -> 44.6 | Improved |
| Add global attention to the basic RNN (ablation) | Stage I duplicate removal · COCO val, FPN | mAP: 34.3 -> 36.7 | Improved |
| Add context gate to the basic RNN (ablation) | Stage I duplicate removal · COCO val, FPN | mAP: 34.3 -> 35.1 | Improved |
| Add global attention and context gate to the basic RNN (ablation) | Stage I duplicate removal · COCO val, FPN | mAP: 34.3 -> 37.2 | Improved |
| Add global attention and context gate to the basic RNN (ablation) | Stage I duplicate removal · COCO val, Mask R-CNN | mAP: 35.9 -> 39.1 | Improved |
| Add global attention and context gate to the basic RNN (ablation) | Stage I duplicate removal · COCO val, PANet with DCN | mAP: 40.6 -> 43.8 | Improved |
| Order proposals by descending score (ablation) | Sequential context encoding · COCO val, FPN | mAP: 33.8 -> 38.3 | Improved |
| Add appearance feature f_A (ablation) | Sequential context encoding · COCO val, FPN | mAP: 35.4 -> 38.3 | Improved |
| Add rank feature f_R (ablation) | Sequential context encoding · COCO val, FPN | mAP: 37.6 -> 38.3 | Improved |
| Add geometric box feature f_G (ablation) | Sequential context encoding · COCO val, FPN | mAP: 37.5 -> 38.3 | Improved |
Closest papers by meaning
- Order Matters: Competition-Guided Query Ordering for RNN-Based Object Detection (NeurIPS 2026 · similarity 0.85)
- ProposalCLIP: Unsupervised Open-Category Object Proposal Generation via Exploiting CLIP Cues (CVPR 2022 · similarity 0.84)
- Should All Proposals Be Treated Equally in Object Detection? (ECCV 2022 · similarity 0.84)
- Spatial Memory for Context Reasoning in Object Detection (ICCV 2017 · similarity 0.84)
- Referring Expression Comprehension Using Language Adaptive Inference (AAAI 2023 · similarity 0.84)
- Hybrid Proposal Refiner: Revisiting DETR Series from the Faster R-CNN Perspective (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.
