Research map / Detection and segmentation / Open-set and few-shot detection
nocaps: novel object captioning at scale
Harsh Agrawal, Karan Desai, Yufei Wang, Xinlei Chen, Rishabh Jain, Mark Johnson, Dhruv Batra, Devi Parikh, Stefan Lee, Peter Anderson
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 constrained beam search with Open Images detections | UpDown · nocaps test (out-of-domain) | CIDEr: 30.1 -> 66.7 | Improved |
| add constrained beam search | UpDown · nocaps val (overall) | CIDEr: 55.3 -> 73.1 | Improved |
| add constrained beam search | UpDown · COCO val 2017 | CIDEr: 116.2 -> 97.7 | Hurt |
| add ELMo language model to CBS | UpDown + CBS · nocaps val (out-of-domain) | CIDEr: 66.4 -> 71.7 | Improved |
| add ground-truth object detections | UpDown + ELMo + CBS · nocaps val (overall) | CIDEr: 74.3 -> 83.3 | Improved |
| add CBS | NBT · nocaps val (overall) | CIDEr: 53.9 -> 61.9 | Improved |
| add CBS | NBT · COCO val 2017 | CIDEr: 88.3 -> 80.2 | Hurt |
| add class-based filtering of CBS constraints (ablation) | UpDown + CBS · nocaps val (overall) | CIDEr: 73.1 -> 73.1 | No clear effect |
| add both class and overlap filtering (ablation) | UpDown + CBS · nocaps val (overall) | CIDEr: 68.2 -> 73.1 | Improved |
| add both filters (ablation) | UpDown + ELMo + CBS · nocaps val (overall) | CIDEr: 69.6 -> 74.3 | Improved |
| add overlap filtering (ablation) | UpDown + ELMo + CBS · nocaps val (overall) | CIDEr: 74.2 -> 74.3 | No clear effect |
| images farther from COCO (decile 10 vs decile 1) | UpDown · nocaps test CIDEr by decile | CIDEr: 82.6 -> 28.7 | Hurt |
Closest papers by meaning
- CAPability: A Comprehensive Visual Caption Benchmark for Evaluating Both Correctness and Thoroughness (NeurIPS 2025 · similarity 0.89)
- Benchmarking Large Vision-Language Models via Directed Scene Graph for Comprehensive Image Captioning (CVPR 2025 · similarity 0.89)
- DetCLIPv3: Towards Versatile Generative Open-vocabulary Object Detection (CVPR 2024 · similarity 0.89)
- LLMDet: Learning Strong Open-Vocabulary Object Detectors under the Supervision of Large Language Models (CVPR 2025 · similarity 0.88)
- CapsFusion: Rethinking Image-Text Data at Scale (CVPR 2024 · similarity 0.88)
- Segment and Caption Anything (CVPR 2024 · similarity 0.88)
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.
