{
  "video": "video-aa5de6b4.mp4",
  "description": "The video appears to be a presentation or demonstration of a research project titled **\"HandX: Scaling Bimanual Motion and Interaction Generation.\"**\n\nHere is a detailed breakdown of what is visible in the video frames:\n\n**00:00 - 00:01 (Introduction/Title Slide):**\n*   The main title is **\"HandX: Scaling Bimanual Motion and Interaction Generation.\"**\n*   A list of authors is present, including Zimu Zhang, Yucheng Zhang, Xiyan Xu, Ziyi Wang, Sirui Xu, Kai Zhou, Bing Zhou, Chuan Guo, Jian Wang, Yu-Xiong Wang, and Liang Yan Gu.\n*   Affiliations mention the \"University of Illinois Urbana-Champaign\" and \"Sense Inc.\"\n*   A citation reference, **\"CVPR 2026\"**, is displayed.\n*   The video features a visual segment showing human hands interacting with small objects (possibly beads or small toys) in a simulation or captured environment, suggesting the focus is on hand motion and manipulation.\n\n**00:01 - 00:04 (Dataset Introduction):**\n*   The focus shifts to a presentation slide introducing the **\"HandX Dataset.\"**\n*   Key statistics about the dataset are prominently displayed:\n    *   **54.2 Hours** (of video data, presumably)\n    *   **5.9M Frames**\n    *   **490K Text** (suggesting associated descriptive data or annotations)\n*   This section emphasizes the scale and richness of the data collected for the HandX project.\n\n**In summary:**\nThe video introduces a sophisticated research effort named **HandX**. The goal of this project is to scale the generation of complex tasks involving two hands (bimanual motion) and interactions with objects. The centerpiece of the demonstration is the **HandX Dataset**, which is massive, containing over 54 hours of high-fidelity video data. The initial visual snippets show realistic representations of hands performing intricate manipulations.",
  "codec": "av1",
  "transcoded": true,
  "elapsed_s": 10.4
}