u-net-brain-tumor icon indicating copy to clipboard operation
u-net-brain-tumor copied to clipboard

View brain image and segmented image.

Open adarsheng1 opened this issue 5 years ago • 4 comments

How we can view the brain MRI and finally the segmented image.

adarsheng1 avatar Apr 09 '19 14:04 adarsheng1

https://github.com/Lafite-Yu/BraTS_2018_U-Net/blob/master/run.py

Lafite-Yu avatar Jun 08 '19 13:06 Lafite-Yu

Thank you Lafite-Yu for the code. I was able to successfully execute it. I have generated the .npz model for enhance and necrotic tumor. How can I know that which type of Tumor it is by executing the code on an input file of type .nii.gz

adarsheng1 avatar Jun 23 '19 09:06 adarsheng1

According to the readme document of this project, you should train one model for one task, relatively. And I didn't change this, you can change the code in train.py or use command like python train.py --task edema to train to separate the edema part of the tumor.

    parser.add_argument('--task', type=str, default='all', help='all, necrotic, edema, enhance')
    # there are 4 labels in targets:
    # Label 0: background
    # Label 1: necrotic and non-enhancing tumor
    # Label 2: edema
    # Label 4: enhancing tumor

You can see know more about that in train.py.

Lafite-Yu avatar Jun 25 '19 07:06 Lafite-Yu

There are four values in the segmentation results of the BraTS dataset, shown as previous comment.

Lafite-Yu avatar Jun 25 '19 07:06 Lafite-Yu