AI Paper Atlas

Research map / Video / Video-language understanding

Temporal and Cross-modal Alignment for Enhanced Audiovisual Video Captioning

Chen Zhao, Jiajun Ma, Qilong Huang, Tiehan Fan, Hongyu Li, Zhuoliang Kang, Xiaoming Wei, Jian Yang, Ying Tai

ECCV 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
fine-tune Qwen3-Omni with OCC-refined data (TCA-Captioner)Qwen3-Omni-30B-A3B · TCA-BenchAV Binding total: 42.5 -> 73.2Improved
fine-tune Qwen3-Omni with OCC-refined data (TCA-Captioner)Qwen3-Omni-30B-A3B · TCA-BenchAV Temporal total: 47.9 -> 76.9Improved
fine-tune Qwen3-Omni with OCC dataQwen3-Omni · video-SALMONN-2Total error: 45.6 -> 31.5Improved
add Audio Checker to Observer data (ablation)OCC pipeline · TCA-BenchAudio: 57.6 -> 61.5Improved
add Visual Checker to Observer data (ablation)OCC pipeline · TCA-BenchVisual: 56.1 -> 64.8Improved
add A/V-Clip-Checker to full OCC (ablation)OCC pipeline · TCA-BenchAV Binding: 66.5 -> 73.2Improved
add A/V-Clip-Checker to full OCC (ablation)OCC pipeline · TCA-BenchAV Temporal: 69.8 -> 76.9Improved
add Audio Checker alone (ablation)OCC pipeline · TCA-BenchAV Temporal: 69.8 -> 68.1Hurt
add DPO with 2,000 OCC preference pairs (ablation)Qwen3-Omni-30B SFT · TCA-BenchAV Temporal: 76.9 -> 79.2Improved
add DPO with OCC preference pairs (ablation)Qwen3-Omni-30B SFT · TCA-BenchVisual: 63.6 -> 66.8Improved
use Qwen2.5-Omni instead of Qwen3-Omni as backbone (ablation)TCA-Captioner SFT · TCA-BenchAV Temporal: 76.9 -> 73.7Hurt
TCA-Captioner vs closed-source Gemini-3.0-ProTCA-Captioner · TCA-BenchAV Binding total: 76.8 (Gemini-3.0-Pro) vs 73.2Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy