New benchmark for search-grounded video misinformation detection

Arxiv CS.CV (computer vision)

Introduces a benchmark where authentic footage is manipulated via editing, reordering, splicing, or AI-generated content to create false narratives. The benchmark focuses on semantic-level misinformation detection.

benchmarks

Arxiv CS.CV (computer vision)·Tao YuYujia YangShenghua ChaiZhang JinshuaiHaopeng JinHao WangMinghui ZhangZhongtian LuoYuchen LongXinlong ChenJiabing YangZhaolu KangYuxuan ZhouZhengyu ManXinming WangHongzhu YiZheqi HeXi YangYan HuangLiang Wang

7 days ago