some strange samples in brats2018 validation dataset

IAmSuyogJadhav / 3d-mri-brain-tumor-segmentation-using-autoencoder-regularization

Keras implementation of the paper "3D MRI brain tumor segmentation using autoencoder regularization" by Myronenko A. (https://arxiv.org/abs/1810.11654).

MIT License

369 stars 114 forks source link

some strange samples in brats2018 validation dataset #28

Closed likeeeee closed 4 years ago

likeeeee commented 4 years ago

Hi! I tested brats2018 validation dataset using my own U-Net models. But I found some very strange samples that the dice score of those is always 0 or 1. I want to ask what features of samples would cause this phenomenon.

IAmSuyogJadhav commented 4 years ago

Hmm. That seems like a severe case of overfitting to me. Closing due to inactivity. Feel free to reopen.