{"ID":6029868,"CreatedAt":"2026-07-08T02:57:47.77373338Z","UpdatedAt":"2026-07-10T16:20:48.775981082Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.06468","arxiv_id":"2607.06468","title":"EgoPolice: A Benchmark for Egocentric Video Understanding in High-Stakes Police Body-Worn Camera Footage","abstract":"We introduce EgoPolice, a carefully curated dataset of real, egocentric police-civilian interactions, sourced from publicly available body-worn camera videos. We select police-civilian action labels that are critical for police behavioral research and annotate them at a second-by-second granularity. The videos feature rapid and irregular camera motion, dense human interactions, and rare high-stakes events, making the dataset a challenging benchmark for motion-robust and context-aware egocentric perception. We provide two different tasks, classification and multiple-choice question-answering, and benchmark both open-source and closed-source models. We find that even the best video models like Gemini 2.5 Pro still struggle to accurately predict high-risk actions such as \"Weapon Out\". Beyond serving as a benchmark, EgoPolice provides a foundation for developing models capable of identifying events of interest in large-scale body-worn camera video repositories, enabling more efficient downstream human review.","short_abstract":"We introduce EgoPolice, a carefully curated dataset of real, egocentric police-civilian interactions, sourced from publicly available body-worn camera videos. We select police-civilian action labels that are critical for police behavioral research and annotate them at a second-by-second granularity. The videos feature...","url_abs":"https://arxiv.org/abs/2607.06468","url_pdf":"https://arxiv.org/pdf/2607.06468v1","authors":"[\"Max Gonzalez Saez-Diez\",\"Jihoon Chung\",\"Adam D. Wolsky\",\"Gregory Lanzalotto\",\"Dean Knox\",\"Jonathan Mummolo\",\"Brandon M. Stewart\",\"Olga Russakovsky\"]","published":"2026-07-07T16:28:10Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false}