Yuhui Ma

@mhkoosheshi Please make sure that you have modified all 512 as 256 in dfn_model.py.

@mhkoosheshi 1. One iter means that you feed one batch of samples into the network during the training phase, while one epoch means that feed the whole training set into...

Thanks for the comments. How many epochs have you trained? And the code is only able to split foreground and background now.

I regret to tell you that my dataset is data of our laboratory, which is not compromised. I am so sorry.

Please make sure all your images have three channels. The code is not perfect. If there are any bugs, please give me more advice.

Yes, because I apply the project on OCT images to complete the task of choroidal segmentation. The code will be updated later to accommodate multi-segmentation tasks.

I am sorry that ImageNet pretrained ResNet101 model hasn't been used in my project. The code needs to be further refined.

Sorry, I didn't notice the problem, and I am not sure whether it could influence the final performance of the network. Perhaps you could use random.seed.

> I solved my problem. Mine is caused by the layer_name. In my model, I use the X101_FPN model from detectron2 (https://github.com/facebookresearch/detectron2/blob/master/configs/COCO-Detection/faster_rcnn_X_101_32x8d_FPN_3x.yaml). The original code of this repo uses the...