arXiv · 机器人与计算机视觉· Ozan Kanbertay (Institute of Computer Science and Campus Institute Data Science, University of G\"ottingen), Richard Vogg (Institute of Computer Science and Campus Institute Data Science, University of G\"ottingen, Behavioral Ecology & Sociobiology Unit, German Primate Center, G\"ottingen, Germany), Elif Karakoc (Behavioral Ecology & Sociobiology Unit, German Primate Center, G\"ottingen, Germany), Peter M. Kappeler (Behavioral Ecology & Sociobiology Unit, German Primate Center, G\"ottingen, Germany, Department of Sociobiology/Anthropology, University of G\"ottingen, G\"ottingen, Germany), Claudia Fichtel (Behavioral Ecology & Sociobiology Unit, German Primate Center, G\"ottingen, Germany), Alexander S. Ecker (Institute of Computer Science and Campus Institute Data Science, University of G\"ottingen)·· 2026-10-20AI 评分4
SiLVi:支持视频交互标注的开源计算机视觉工具
SiLVi: Simple Interface for Labeling Video Interactions
AI 导读
哥廷根大学团队发布开源软件 SiLVi,填补了现有工具在行为标注与个体定位间无法同时实现的空白。该软件允许研究人员直接在视频中注释行为及互动,生成结构化数据以训练和验证计算机视觉模型。其文档、下载说明及 Linux 版本已提供,适用于从动物行为分析到提取动态场景图的人类交互标注任务。
来源:arXiv · 机器人与计算机视觉 · arxiv.org