Video2Script

基本概念

文献调研

• VideoChat

  [2305.06355] VideoChat: Chat-Centric Video Understanding (opens new window)

  【双流：逐帧+视频】

• MVBench（VideoChat2）

  [2311.17005] MVBench: A Comprehensive Multi-modal Video Understanding Benchmark (opens new window)

• Dolphin

  GitHub - kaleido-lab/dolphin: General video interaction platform based on LLMs, including Video ChatGPT (opens new window)

  【开源项目】

相关工作

其他工作

• GRiT

  [2212.00280] GRiT: A Generative Region-to-text Transformer for Object Understanding (arxiv.org) (opens new window)

• Dense Video Object Captioning from Disjoint Supervision

  [2306.11729] Dense Video Object Captioning from Disjoint Supervision (arxiv.org) (opens new window)

• A Challenging Multimodal Video Summary: Simultaneously Extracting and Generating Keyframe-Caption Pairs from Video

  [2312.01575] A Challenging Multimodal Video Summary: Simultaneously Extracting and Generating Keyframe-Caption Pairs from Video (arxiv.org) (opens new window)

• Vid2Seq（CVPR-23）

  [2302.14115] Vid2Seq: Large-Scale Pretraining of a Visual Language Model for Dense Video Captioning (arxiv.org) (opens new window)

• VidChapters-7M（NIPS-23）

  [2309.13952] VidChapters-7M: Video Chapters at Scale (arxiv.org) (opens new window)