{
  "id": 17,
  "url": "https://arxiv.org/abs/2607.11167v1",
  "title": "Pix2Act: Image-Space Manipulation Policies with Equivariant Augmentation",
  "summary": "Representing manipulation actions as 2D trajectories in the camera plane provides a compact and interpretable basis for learning complex 3D manipulation policies. However, it also creates challenges from out-of-frame trajectories and limited precision. We propose Pix2Act, an imitation learning method that addresses these challenges by generating continuous image-space keypoint trajectories in each camera plane and losslessly recovering end-effector poses via triangulation. This reformulates high",
  "authors": "Haojie Huang, Linfeng Zhao, Haotian Liu, Zhang Ye, Si-Yuan Huang, Mingxi Jia et al.",
  "category": "research",
  "topics": null,
  "orgs": null,
  "regions": null,
  "published_at": "2026-07-13T07:03:29.000Z",
  "fetched_at": "2026-07-14T14:14:15.663Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/17",
  "original_url": "https://arxiv.org/abs/2607.11167v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}