How to use deepmask/sharpmask for object localization problem (bounding box) instead of segmentation problem?
yxchng opened this issue · comments
I just need to train a more that has bounding box label and predict the bounding boxes. How should I prepare my data since I don't have segmentation label? Thanks.