Research map / Detection and segmentation / CNN features
Deep Regionlets for Object Detection
Hongyu Xu, Xutao Lv, Xiaoyu Wang, Zhou Ren, Navaneeth Bodla, Rama Chellappa
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 region selection network with multiple regions instead of one global region (ablation) | Deep Regionlets · VOC07 test (ResNet-101) | mAP@0.5 (%): 30.27 -> 82.0 | Improved |
| add non-rectangular region selection over offset-only rectangular regions (ablation) | Deep Regionlets · VOC07 test (ResNet-101) | mAP@0.5 (%): 78.5 -> 81.3 (+2.8) | Improved |
| add gating network for soft regionlet selection (ablation) | Deep Regionlets · VOC07 test (ResNet-101) | mAP@0.5 (%): 81.3 -> 82.0 | Improved |
| increase selected regions from 4 to 9 (4x4 regionlets) (ablation) | Deep Regionlets · VOC07 test (ResNet-101) | mAP@0.5 (%): 79.9 -> 80.9 | Improved |
| increase selected regions from 9 to 16 (4x4 regionlets) (ablation) | Deep Regionlets · VOC07 test (ResNet-101) | mAP@0.5 (%): 80.9 -> 82.0 | Improved |
| increase regionlet density from 2x2 to 4x4 (16 regions) (ablation) | Deep Regionlets · VOC07 test (ResNet-101) | mAP@0.5 (%): 80.0 -> 82.0 | Improved |
| increase regionlet density from 5x5 to 6x6 (16 regions) (ablation) | Deep Regionlets · VOC07 test (ResNet-101) | mAP@0.5 (%): 81.6 -> 80.8 | Hurt |
| apply soft-NMS at test time only | Deep Regionlets (VGG16) · VOC07 test (train 07+12) | mAP@0.5 (%): 79.2 -> 80.1 | Improved |
| apply soft-NMS at test time only | Deep Regionlets (ResNet-101) · VOC07 test | mAP@0.5 (%): 82.0 -> 83.1 | Improved |
| swap VGG16 backbone for ResNet-101 | Deep Regionlets · VOC07 test (train 07+12) | mAP@0.5 (%): 79.2 -> 82.0 | Improved |
| add more training data (07+12 instead of 07) | Deep Regionlets (VGG16) · VOC07 test | mAP@0.5 (%): 73.0 -> 79.2 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- R-FCN — R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016) · better on 3 of 3 shared results
- Faster R-CNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Deep Regionlets: Blended Representation and Deep Learning for Generic Object Detection (TPAMI 2019 · similarity 0.99)
- Region-Based Convolutional Networks for Accurate Object Detection and Segmentation (TPAMI 2016 · similarity 0.87)
- Object Detection Networks on Convolutional Feature Maps (TPAMI 2016 · similarity 0.87)
- R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016 · similarity 0.86)
- Recognize Any Regions (NeurIPS 2024 · similarity 0.86)
- Training-Free Open-Ended Object Detection and Segmentation via Attention as Prompts (NeurIPS 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.
