how to train 'Retina U-Net' on customer dataset

MIC-DKFZ / medicaldetectiontoolkit

The Medical Detection Toolkit contains 2D + 3D implementations of prevalent object detectors such as Mask R-CNN, Retina Net, Retina U-Net, as well as a training and inference framework focused on dealing with medical images.

Apache License 2.0

1.31k stars 297 forks source link

how to train 'Retina U-Net' on customer dataset #78

Closed adeebakausar closed 5 years ago

adeebakausar commented 5 years ago

Explain the 'Folder Structure' of images ,annotation and mask etc. .. Give a proper example with folders structure of images and annotations,

pfjaeger commented 5 years ago

please run the generation script for the toy dataset, it will create a folder structure that should be explicatory.