Research map / Video / Video-language understanding
Leveraging Weighted Cross-Graph Attention for Visual and Semantic Enhanced Video Captioning Network
Deepali Verma, Arya Haldar, Tanima Dutta
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 graph construction to the encoder-decoder baseline (ablation) | VSVCap · YouTube2Text | BLEU@4: 42.7 -> 57.3 | Improved |
| add fine-grained cross-graph attention after graph construction (ablation) | VSVCap · YouTube2Text | BLEU@4: 57.3 -> 58.6 | Improved |
| add visual-semantic reasoning (ablation) | VSVCap · YouTube2Text | BLEU@4: 58.6 -> 60.9 | Improved |
| add graph construction to the encoder-decoder baseline (ablation) | VSVCap · MSR-VTT | BLEU@4: 30.6 -> 42.9 | Improved |
| add fine-grained cross-graph attention after graph construction (ablation) | VSVCap · MSR-VTT | BLEU@4: 42.9 -> 44.3 | Improved |
| add visual-semantic reasoning (ablation) | VSVCap · MSR-VTT | BLEU@4: 44.3 -> 46.1 | Improved |
| use a directed rather than undirected weighted semantic graph (ablation) | VSVCap · YouTube2Text | BLEU@4: 58.3 -> 60.9 | Improved |
| use a directed rather than undirected weighted semantic graph (ablation) | VSVCap · MSR-VTT | BLEU@4: 44.1 -> 46.1 | Improved |
| vary the GCN kernel count from 2 to 8 (ablation) | VSVCap · YouTube2Text | BLEU@4: 55.0 -> 60.9 | Improved |
| vary the GCN kernel count from 8 to 16 (ablation) | VSVCap · MSR-VTT | BLEU@4: 46.1 -> 46.2 | No clear effect |
| use VSVCap instead of Seq2Seq | video captioning · Human evaluation | relevancy: 0.5043 -> 0.9095 | Improved |
Closest papers by meaning
- Fine-Grained Captioning of Long Videos through Scene Graph Consolidation (ICML 2025 · similarity 0.89)
- AuroraCap: Efficient, Performant Video Detailed Captioning and a New Benchmark (ICLR 2025 · similarity 0.87)
- Exploiting Auxiliary Caption for Video Grounding (AAAI 2024 · similarity 0.86)
- Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers (CVPR 2024 · similarity 0.86)
- Progress-Aware Video Frame Captioning (CVPR 2025 · similarity 0.86)
- Distilling Vision-Language Models on Millions of Videos (CVPR 2024 · similarity 0.86)
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.
