关于preprocess的遇到一点问题 about anatomynet-for-anatomical-segmentation HOT 11 CLOSED

For .npy, it should be read by numpy, not simpleitk.

from anatomynet-for-anatomical-segmentation.

from anatomynet-for-anatomical-segmentation.

Find the line, then change to np.read()

from anatomynet-for-anatomical-segmentation.

from anatomynet-for-anatomical-segmentation.

We use three datasets. You can get our collected three datasets from https://github.com/wentaozhu/AnatomyNet-for-anatomical-segmentation/blob/master/README.md
There are dataset paths.

from anatomynet-for-anatomical-segmentation.

from anatomynet-for-anatomical-segmentation.

Try to format it as the these datasets.

from anatomynet-for-anatomical-segmentation.

from anatomynet-for-anatomical-segmentation.

We use an indicator vector flagvect in https://github.com/wentaozhu/AnatomyNet-for-anatomical-segmentation/blob/master/src/AnatomyNet.py#L225

from anatomynet-for-anatomical-segmentation.

from anatomynet-for-anatomical-segmentation.

It is generated automatically if you run the script. You can get the re-implementation from https://github.com/Project-MONAI/MONAI/tree/master/research/lamp-automated-model-parallelism

from anatomynet-for-anatomical-segmentation.