|
a/README.md |
|
b/README.md |
1 |
### Quick Links |
1 |
### Quick Links |
2 |
|
2 |
|
3 |
- [Development Branch](https://github.com/GalDude33/Fetal-MRI-Segmentation/tree/dev) |
3 |
- [Development Branch](https://github.com/GalDude33/Fetal-MRI-Segmentation/tree/dev) |
4 |
|
4 |
|
5 |
- [Experiments Branch](https://github.com/GalDude33/Fetal-MRI-Segmentation/tree/adapt) |
5 |
- [Experiments Branch](https://github.com/GalDude33/Fetal-MRI-Segmentation/tree/adapt) |
6 |
|
6 |
|
7 |
- [Augmentations Notebook](https://github.com/GalDude33/Fetal-MRI-Segmentation/blob/adapt/notebooks/Check_Augmentations.ipynb) |
7 |
- [Augmentations Notebook](https://github.com/GalDude33/Fetal-MRI-Segmentation/blob/adapt/notebooks/Check_Augmentations.ipynb) |
8 |
|
8 |
|
9 |
### Fetal Total-Body and Brain 3D Segmentation in MRI scans with limited datasets. |
9 |
### Fetal Total-Body and Brain 3D Segmentation in MRI scans with limited datasets. |
10 |
|
10 |
|
11 |
- 3D and 2D segmentation. |
11 |
- 3D and 2D segmentation.
|
12 |
- Implemented Advanced Augmentation techniques. |
12 |
- Implemented Advanced Augmentation techniques.
|
13 |
- Achieving results below Radiologist Inter & Intra Observer variability. |
13 |
- Achieving results below Radiologist Inter & Intra Observer variability.
|
14 |
- Option to Utilize previous-slice segmentation to improve performance on hard samples. |
14 |
- Option to Utilize previous-slice segmentation to improve performance on hard samples. |
15 |
|
15 |
|
16 |
### Training |
16 |
### Training |
17 |
|
17 |
|
18 |
* Install Tensorflow & Keras |
18 |
* Install Tensorflow & Keras |
19 |
|
19 |
|
20 |
* Install dependencies: |
20 |
* Install dependencies:
|
21 |
pip install -r requirements.txt |
21 |
pip install -r requirements.txt |
22 |
|
22 |
|
23 |
* To run training: |
23 |
* To run training:
|
24 |
Config currently in dict inside brats/train_fetal.py |
24 |
Config currently in dict inside brats/train_fetal.py
|
25 |
``` |
25 |
```
|
26 |
$ python -m brats.train_fetal --config_dir <config_dir> |
26 |
$ python -m brats.train_fetal --config_dir <config_dir>
|
27 |
``` |
27 |
```
|
28 |
Where <config_dir> is the folder containing the training configurations (or will be creatad to) |
28 |
Where <config_dir> is the folder containing the training configurations (or will be creatad to) |
29 |
|
29 |
|
30 |
### Write prediction images from the validation data |
30 |
### Write prediction images from the validation data
|
31 |
In the training above, part of the data was held out for validation purposes. |
31 |
In the training above, part of the data was held out for validation purposes.
|
32 |
To write the predicted label maps to file: |
32 |
To write the predicted label maps to file:
|
33 |
``` |
33 |
```
|
34 |
$ python -m brats.predict --config_dir <config_dir> |
34 |
$ python -m brats.predict --config_dir <config_dir>
|
35 |
``` |
35 |
```
|
36 |
The predictions will be written in the ```<config_dir>/prediction``` folder along with the input data and ground truth labels for comparison. |
36 |
The predictions will be written in the ```<config_dir>/prediction``` folder along with the input data and ground truth labels for comparison. |
37 |
|
37 |
|
38 |
### More Instructions and Explanations in the [Wiki](../../wiki)! |
38 |
### More Instructions and Explanations in the [Wiki](../../wiki)! |
39 |
|
39 |
|
40 |
### Cool Results |
40 |
### Cool Results |
41 |
|
41 |
|
42 |
##### 3D Model of the segmented fetal brain |
42 |
##### 3D Model of the segmented fetal brain
|
43 |
 |
43 |

|
44 |
##### 3D Model of the segmented fetal body |
44 |
##### 3D Model of the segmented fetal body
|
45 |
 |
45 |

|
46 |
##### Segmentation of fetal brain |
46 |
##### Segmentation of fetal brain
|
47 |
 |
47 |

|
48 |
##### Segmentation of fetal brain (only border) |
48 |
##### Segmentation of fetal brain (only border)
|
49 |
 |
49 |

|
50 |
##### Segmentation of fetal body |
50 |
##### Segmentation of fetal body
|
51 |
 |
51 |

|