Research map / Video / Video-language understanding
Controllable Video Captioning With POS Sequence Guidance Based on Gated Fusion Network
Bairui Wang, Lin Ma, Wei Zhang, Wenhao Jiang, Jingwen Wang, Wei Liu
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 cross gating fusion instead of concatenation (ablation) | EncDec (IR+M) · MSR-VTT test | CIDEr: 45.4 -> 48.4 | Improved |
| add cross gating fusion instead of concatenation (ablation) | EncDec (I3D+M) · MSR-VTT test | CIDEr: 45.1 -> 47.7 | Improved |
| add cross gating fusion (ablation) | EncDec (IR+M) · MSR-VTT test | BLEU@4: 39.7 -> 41.7 | Improved |
| add POS sequence guidance (ablation) | EncDec+CG (IR+M) · MSR-VTT test | CIDEr: 48.4 -> 48.7 | Improved |
| add POS sequence guidance (ablation) | EncDec+CG (I3D+M) · MSR-VTT test | CIDEr: 47.7 -> 48.5 | Improved |
| add POS sequence guidance (ablation) | EncDec+CG (IR+M) · MSR-VTT test | BLEU@4: 41.7 -> 42.0 | Improved |
| use cross gating instead of MCB fusion (ablation) | EncDec (IR+M) · MSR-VTT test | CIDEr: 46.3 -> 48.4 | Improved |
| use cross gating instead of MLB fusion (ablation) | EncDec (IR+M) · MSR-VTT test | CIDEr: 47.6 -> 48.4 | Improved |
| use cross gating instead of element-wise adding (ablation) | EncDec (IR+M) · MSR-VTT test | CIDEr: 46.3 -> 48.4 | Improved |
| train with self-critical RL instead of cross-entropy (ablation) | Ours (IR+M) · MSR-VTT test | CIDEr: 48.7 -> 53.4 | Improved |
| train with self-critical RL instead of cross-entropy | Ours (IR+M) · MSR-VTT test | BLEU@4: 42.0 -> 41.3 | Hurt |
| train with self-critical RL instead of cross-entropy | Ours (IR+M) · MSVD test | CIDEr: 88.7 -> 91.0 | Improved |
Closest papers by meaning
- Bidirectional Attentive Fusion With Context Gating for Dense Video Captioning (CVPR 2018 · similarity 0.86)
- Joint Syntax Representation Learning and Visual Cue Translation for Video Captioning (ICCV 2019 · similarity 0.85)
- Spatio-Temporal Dynamics and Semantic Attribute Enriched Visual Encoding for Video Captioning (CVPR 2019 · similarity 0.85)
- Parallelized Autoregressive Decoding for Omni-Modal Dense Video Captioning (ECCV 2026 · similarity 0.85)
- Distilling Vision-Language Models on Millions of Videos (CVPR 2024 · similarity 0.84)
- InstanceCap: Improving Text-to-Video Generation via Instance-aware Structured Caption (CVPR 2025 · similarity 0.83)
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.
