VISION Datasets
收藏资源简介:
VISION Datasets是由佐治亚理工学院等机构创建的一套包含14个工业检测数据集的集合,旨在解决工业视觉检测中的数据可用性、质量和复杂生产需求等问题。该数据集包含18,000张图像,涵盖44种缺陷类型,支持多种检测方法,如无监督、弱监督、半监督和监督缺陷检测。数据集通过实例分割标注,精确识别缺陷,适用于多种制造过程、材料和行业。此外,VISION Datasets还支持两项挑战竞赛,旨在推动工业视觉检测技术的进步,解决实际生产中的复杂问题。
VISION Datasets is a collection of 14 industrial inspection datasets created by institutions such as the Georgia Institute of Technology, with the goal of addressing challenges including data availability, quality issues, and complex production requirements in industrial visual inspection. This collection comprises 18,000 images spanning 44 defect categories, and supports a variety of inspection approaches, namely unsupervised, weakly-supervised, semi-supervised, and supervised defect detection. Annotated through instance segmentation, the datasets allow for accurate identification of defects, making them applicable across diverse manufacturing processes, materials, and industrial sectors. Furthermore, VISION Datasets also features two challenge competitions, intended to promote the advancement of industrial visual inspection technologies and resolve complex practical problems in production.

- 1VISION Datasets: A Benchmark for Vision-based InduStrial InspectiON佐治亚理工学院 · 2023年



