Research map / Detection and segmentation / CNN features
Objects365: A Large-Scale, High-Quality Dataset for Object Detection
Shuai Shao, Zeming Li, Tianyuan Zhang, Chao Peng, Gang Yu, Xiangyu Zhang, Jing Li, Jian Sun
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 |
|---|---|---|---|
| pretrain on Objects365 instead of ImageNet | FPN · COCO val (90K finetune) | mmAP: 42 vs 36.4 | Improved |
| pretrain on Objects365 instead of ImageNet | RetinaNet · COCO val (180K finetune) | mmAP: 39.5 vs 37 | Improved |
| use L2-norm-scaled smaller learning rate for Objects365 finetuning (ablation) | FPN Objects365 pretrain · COCO val | mmAP: 40.4 -> 42.0 | Improved |
| pretrain on Objects365 instead of ImageNet | FPN · PASCAL VOC detection | mAP: 86.2 vs 80.2 | Improved |
| pretrain on Objects365 instead of ImageNet | FPN · CityPersons | MR: 12.12 vs 18.04 | Improved |
| pretrain on Objects365 instead of ImageNet | PSPNet · PASCAL VOC segmentation | mIOU: 76.7 vs 74.5 | Improved |
| pretrain on Objects365 instead of ImageNet | PSPNet · ADE20K val | mIOU: 42.9 vs 41.2 | Improved |
| add ImageNet pretraining before Objects365 (ablation) | FPN · COCO val (90K) | mmAP: 42.3 vs 42 | No clear effect |
| add ImageNet pretraining before Objects365 (ablation) | RetinaNet · COCO val | mmAP: 41 vs 39.5 | Improved |
| pretrain on Objects365 including the 80 COCO categories (ablation) | FPN · COCO val (90K) | mmAP: 39.6 -> 42.0 | Improved |
| pretrain the head as well as the backbone (ablation) | FPN Objects365 · COCO val | mmAP: 42.0 vs 39.4 | Improved |
| add ImageNet pretraining before Objects365 (ablation) | PSPNet · ADE20K val | mIOU: 42.9 -> 43.3 | Improved |
Closest papers by meaning
- "A Simple Approach and Benchmark for 21,000-Category Object Detection" (ECCV 2022 · similarity 0.91)
- Towards Large-Scale Small Object Detection: Survey and Benchmarks (TPAMI 2023 · similarity 0.89)
- Region-Based Convolutional Networks for Accurate Object Detection and Segmentation (TPAMI 2016 · similarity 0.89)
- Object Detection Benchmarks are Incomplete: The Role of Label Errors and Annotation Uncertainty (NeurIPS 2026 · similarity 0.88)
- Towards RAW Object Detection in Diverse Conditions (CVPR 2025 · similarity 0.88)
- Roboflow100-VL: A Multi-Domain Object Detection Benchmark for Vision-Language Models (NeurIPS 2025 · similarity 0.87)
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.
