/polygons-vs-bbox

Here I experimented whether is better to directly train an object detection model on bounding boxes or train it on polygons and then convert the prediction to bounding boxes.

Primary LanguageJupyter Notebook

Stargazers