Open LiwenWang919 opened 8 months ago
and work for object detection
Hi, since object detection is quite different from landmark detection, I'm afraid you'll need to change a lot to adapt. It might be easy to take a look at other object detection framework.
Hi, since object detection is quite different from landmark detection, I'm afraid you'll need to change a lot to adapt. It might be easy to take a look at other object detection framework.
Hi, I'm facing 2 problems: 1. How can I get the HeadNew datasets,author did not reply to my emails.2. In my experiment on the chest, the accuracy could not reach the accuracy of the given weights. Is there any hyperparameter that needs to be set?
Hi, since object detection is quite different from landmark detection, I'm afraid you'll need to change a lot to adapt. It might be easy to take a look at other object detection framework.
Hi, I'm facing 2 problems: 1. How can I get the HeadNew datasets,author did not reply to my emails.2. In my experiment on the chest, the accuracy could not reach the accuracy of the given weights. Is there any hyperparameter that needs to be set?
Hi, 1) Since I do not own the HeadNew dataset, there is nothing I can do if the authors do not reply. 2) You may need to tune cls_loss_weight
and domain_loss_weight
as they could be sensitive.
if i want to applicate this model on other medical image dataset,how can i do?