PrepareDataset.ipynb
收藏Recherche Data Gouv France2025-01-01 更新2026-04-09 收录
下载链接：
https://entrepot.recherche.data.gouv.fr/file.xhtml?persistentId=doi:10.57745/FEM9N4
下载链接
链接失效反馈官方服务：
资源简介：
The "Prepare Dataset" Jupyter notebook is used to assemble the training dataset for object detection. It handles the merging and formatting of annotated images, which include polygonal labels for multiple classes (labels, data matrix, leaves, necrosis, and plug), initially created using Segment Anything 2.0 in CVAT. Annotations are exported in LabelMe 3.0 format and converted to COCO format using the datumaro library before being merged into a single dataset suitable for training with Mask R-CNN.
创建时间：
2025-01-01
