Research map / Video / Video-language understanding
It’s Just Another Day: Unique Video Captioning by Discriminitive Prompting
Toby Perrett, Tengda Han, Dima Damen, Andrew Zisserman
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 discriminative prompting (CDP) | LaViLa VCLM · Egocentric benchmark T=+0s | Avg R@1: 37 -> 45 | Improved |
| add discriminative prompting (CDP) with next 5s clip | LaViLa VCLM · Egocentric benchmark T=+5s | Cycle@1: 23.0 -> 38.6 | Improved |
| add discriminative prompting (CDP) with 30s access | LaViLa VCLM · Egocentric benchmark T=+30s | Cycle@1: 27.2 -> 62.3 | Improved |
| add discriminative prompting (CDP) | Video-LLaMA · Timeloop movies T=0s | Cycle@1: 18.3 -> 25.0 | Improved |
| add discriminative prompting (CDP) with next two clips | Video-LLaMA · Timeloop movies T=4s | Cycle@1: 18.4 -> 37.4 | Improved |
| add discriminative prompting (CDP) with 10s access | Video-LLaMA · Timeloop movies T=10s | Avg R@1: 38 -> 63 | Improved |
| add discriminative prompting (CDP) | Video-LLaMA · Timeloop movies T=2s | Text-to-video R@1: 55 -> 51 | Hurt |
| add discriminative prompting (CDP) | Video-LLaMA in CLIP space · Timeloop movies T=0 | Avg R@1: 31 -> 28 | Hurt |
| add discriminative prompting (CDP) to long egocentric videos | LaViLa VCLM · Long egocentric videos | Text-to-video R@1: 12 -> 32 | Improved |
| add CDP to a weaker base captioner | EILEV · Egocentric benchmark T=0 (LaViLa space) | Avg R@1: 15 -> 17 | Improved |
| add CDP to a stronger base captioner | TSF-L LaViLa VCLM · Egocentric benchmark T=30 (EgoVLP space) | Avg R@1: 43 -> 76 | Improved |
| use CDPNet instead of exhaustive caption search (ablation) | CDP · Egocentric captioning time | seconds per captioning: 300 -> 5.8 | Improved |
Closest papers by meaning
- InstanceCap: Improving Text-to-Video Generation via Instance-aware Structured Caption (CVPR 2025 · similarity 0.87)
- AuroraCap: Efficient, Performant Video Detailed Captioning and a New Benchmark (ICLR 2025 · similarity 0.87)
- Video ReCap: Recursive Captioning of Hour-Long Videos (CVPR 2024 · similarity 0.86)
- Progress-Aware Video Frame Captioning (CVPR 2025 · similarity 0.86)
- Set Prediction Guided by Semantic Concepts for Diverse Video Captioning (AAAI 2024 · similarity 0.86)
- Vript: A Video Is Worth Thousands of Words (NeurIPS 2024 · similarity 0.85)
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.
