Camouflaged Object Plus Plus (CAMO++)
收藏资源简介:
Camouflaged Object Plus Plus (CAMO++) 数据集是由越南国家大学胡志明市分校创建的,旨在推动野外伪装实例分割的新任务。该数据集扩展了初步的CAMO数据集,增加了图像数量和任务多样性,包含5500张图像,涵盖超过90种动物物种和人类,所有图像均具有层次像素级标注。CAMO++数据集不仅用于伪装实例分割任务,还用于传统伪装对象分割任务,旨在解决计算机视觉中伪装对象的识别和分割问题。
Camouflaged Object Plus Plus (CAMO++) dataset was developed by Vietnam National University Ho Chi Minh City Campus to advance the novel task of in-the-wild camouflaged instance segmentation. This dataset expands the initial CAMO dataset by increasing the number of images and enriching task diversity. It contains 5500 images covering over 90 animal species and humans, with hierarchical pixel-level annotations for all samples. The CAMO++ dataset supports both camouflaged instance segmentation and traditional camouflaged object segmentation tasks, aiming to address the recognition and segmentation challenges of camouflaged objects in computer vision.

- 1Camouflaged Instance Segmentation In-The-Wild: Dataset, Method, and Benchmark Suite越南国家大学胡志明市分校 · 2021年



