WS2 (Weakly Supervised segmentation for Waste-Sorting)
收藏资源简介:
WS2数据集是第一个多视图数据集,由超过11000个高分辨率视频帧组成,这些帧是在传送带上捕获的,包括“之前”和“之后”的图像。该数据集由意大利Verderio的一个真实世界塑料分类设施收集,用于支持弱监督分割研究。数据集包含一个序列,其中包括在人类操作员从传送带上选择半透明彩色PET物体之前和之后的视频帧。该数据集旨在促进研究,以开发基于深度学习的自动化分类和质量控制解决方案。
The WS2 dataset is the first multi-view dataset composed of over 11,000 high-resolution video frames captured on a conveyor belt, including "before" and "after" images. This dataset was collected from a real-world plastic sorting facility in Verderio, Italy, to support weakly supervised segmentation research. It encompasses a sequence of video frames captured both before and after a human operator selects translucent colored PET objects from the conveyor belt. This dataset is designed to advance research on developing deep learning-based automated classification and quality control solutions.




