Research map / Vision-language and multimodal / Cross-modal learning
CineSRD: Leveraging Visual, Acoustic, and Linguistic Cues for Open-World Visual Media Speaker Diarization
Liangbin Huang, Xiaohua Liao, Chaoqun Cui, Shijing Wang, Zhaolong Huang, Yanlong Du, Wenji Mao
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 visual anchor clustering (audio to audio-visual) (ablation) | CineSRD (SC) · SubtitleSD Chinese | DER: 0.14945 -> 0.08325 | Improved |
| add visual anchor clustering (audio to audio-visual) (ablation) | CineSRD (AHC) · SubtitleSD Chinese-Hard | DER: 0.20679 -> 0.10585 | Improved |
| add text modality via speaker turn detection (AV to AVT) (ablation) | CineSRD (SC) · SubtitleSD Chinese | DER: 0.08325 -> 0.07561 | Improved |
| add text modality via speaker turn detection (AV to AVT) (ablation) | CineSRD (SC) · SubtitleSD Chinese-Hard | JER: 0.55392 -> 0.41462 | Improved |
| add text modality via speaker turn detection (AV to AVT) (ablation) | CineSRD (SC) · SubtitleSD English | DER: 0.10272 -> 0.08932 | Improved |
| add text modality via speaker turn detection (AV to AVT) (ablation) | CineSRD (AHC) · SubtitleSD Chinese | DER: 0.08528 -> 0.07700 | Improved |
| add timbre similarity fusion to ALM (ablation) | ALM Only · SubtitleSD Chinese (turn detection) | F1: 0.83015 -> 0.92715 | Improved |
| add ALM text/audio fusion to timbre similarity (ablation) | Audio Only · SubtitleSD Chinese (turn detection) | F1: 0.86490 -> 0.92715 | Improved |
| add ALM fusion to timbre similarity (ablation) | Audio Only · SubtitleSD Chinese (turn detection) | AUC: 0.95862 -> 0.96257 | Improved |
| use ALM alone instead of timbre similarity alone (ablation) | Speaker turn detection · SubtitleSD Chinese | AUC: 0.95862 -> 0.78974 | Hurt |
| use CineSRD instead of EC2P | EC2P · AVA-AVD | DER: 0.2032 -> 0.1898 | Improved |
| use CineSRD (AVT) instead of EC2P (AVT) | EC2P · SubtitleSD Chinese | DER: 0.13451 -> 0.07561 | Improved |
Closest papers by meaning
- Seeing Speech and Sound: Distinguishing and Locating Audio Sources in Visual Scenes (CVPR 2025 · similarity 0.86)
- Multi-speaker Attention Alignment for Multimodal Social Interaction (CVPR 2026 · similarity 0.86)
- JointAVBench: A Benchmark for Joint Audio-Visual Reasoning Evaluation (ICLR 2026 · similarity 0.86)
- MMOU: A Massive Multi-Task Omni Understanding and Reasoning Benchmark for Long and Complex Real-World Videos (NeurIPS 2026 · similarity 0.85)
- ViLMA: A Zero-Shot Benchmark for Linguistic and Temporal Grounding in Video-Language Models (ICLR 2024 · similarity 0.85)
- OmniCapBench: A Deep-Structured Evaluation Framework for Fine-Grained Audio-Visual Captioning (NeurIPS 2026 · 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.
