Objects and Distractions Segmentation Dataset

The Objects and Distractions Segmentation Dataset is designed for robotics and visual entertainment sectors, featuring a range of internet-collected images with resolutions between 1365 x 2047 and 4165 x 2737 pixels. This dataset emphasizes semantic segmentation, categorizing images into five main types of interference objects, including target persons, objects, interference items, and various human body parts, facilitating the development of algorithms to distinguish between primary subjects and background distractions.

《物体与干扰物分割数据集》专为机器人技术和视觉娱乐行业设计，汇集了1365 x 2047至4165 x 2737像素分辨率的互联网收集图像。本数据集侧重于语义分割，将图像划分为五大类干扰物体，包括目标人物、物体、干扰物品以及多种人体部位，有助于算法的开发，以区分主要主体与背景干扰。