Skip to content

Commit 6355dad

Browse files
jtetreapre-commit-ci[bot]
jtetrea
and
pre-commit-ci[bot]
authored
update prostate lesion (#398)
Updated prostate-mri-lesion-seg to use MONAI Deploy v2.0.0 based on NVIDIA Holoscan. Also added PI-RADS classifier model to pipeline. --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
1 parent d4fa3dc commit 6355dad

21 files changed

Lines changed: 2308 additions & 809 deletions

‎prostate-mri-lesion-seg/LICENSE‎

Lines changed: 54 additions & 54 deletions
Original file line numberDiff line numberDiff line change
@@ -1,54 +1,54 @@
1-
Prostate-MRI_Lesion_Detection, v2.0 (Release date: August 2, 2023)
2-
• DEFINITIONS: AUTHOR(S) NVIDIA Corp. and National Cancer Institute, NIH
3-
4-
• PROVIDER: the National Cancer Institute (NCI), a participating institute of the
5-
National Institutes of Health (NIH), and an agency of the United States Government.
6-
7-
• SOFTWARE: the machine readable, binary, object code form,
8-
and the related documentation for the modules of the Prostate-MRI_Lesion_Detection, v2.0
9-
software package, which is a collection of operators which accept (T2, ADC, and High
10-
b-value DICOM images) and produce prostate organ and lesion segmentation files
11-
12-
• RECIPIENT: the party that downloads the software.
13-
14-
By downloading or otherwise receiving the SOFTWARE, RECIPIENT may
15-
use and/or redistribute the SOFTWARE, with or without modification,
16-
subject to RECIPIENT’s agreement to the following terms:
17-
18-
1. THE SOFTWARE SHALL NOT BE USED IN THE TREATMENT OR DIAGNOSIS
19-
OF HUMAN SUBJECTS. RECIPIENT is responsible for
20-
compliance with all laws and regulations applicable to the use
21-
of the SOFTWARE.
22-
23-
2. THE SOFTWARE is distributed for NON-COMMERCIAL RESEARCH PURPOSES ONLY. RECIPIENT is
24-
responsible for appropriate-use compliance.
25-
26-
3. RECIPIENT agrees to acknowledge PROVIDER’s contribution and
27-
the name of the author of the SOFTWARE in all written publications
28-
containing any data or information regarding or resulting from use
29-
of the SOFTWARE.
30-
31-
4. THE SOFTWARE IS PROVIDED "AS IS" AND ANY EXPRESS OR IMPLIED
32-
WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
33-
OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT
34-
ARE DISCLAIMED. IN NO EVENT SHALL THE PROVIDER OR THE INDIVIDUAL DEVELOPERS
35-
BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
36-
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
37-
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
38-
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
39-
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
40-
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
41-
THE POSSIBILITY OF SUCH DAMAGE.
42-
43-
5. RECIPIENT agrees not to use any trademarks, service marks, trade names,
44-
logos or product names of NVIDIA, NCI or NIH to endorse or promote products derived
45-
from the SOFTWARE without specific, prior and written permission.
46-
47-
6. For sake of clarity, and not by way of limitation, RECIPIENT may add its
48-
own copyright statement to its modifications or derivative works of the SOFTWARE
49-
and may provide additional or different license terms and conditions in its
50-
sublicenses of modifications or derivative works of the SOFTWARE provided that
51-
RECIPIENT’s use, reproduction, and distribution of the SOFTWARE otherwise complies
52-
with the conditions stated in this Agreement. Whenever Recipient distributes or
53-
redistributes the SOFTWARE, a copy of this Agreement must be included with
54-
each copy of the SOFTWARE.
1+
Prostate-MRI_Lesion_Detection, v3.0 (Release date: September 17, 2024)
2+
DEFINITIONS: AUTHOR(S) NVIDIA Corp. and National Cancer Institute, NIH
3+
4+
PROVIDER: the National Cancer Institute (NCI), a participating institute of the
5+
National Institutes of Health (NIH), and an agency of the United States Government.
6+
7+
SOFTWARE: the machine readable, binary, object code form,
8+
and the related documentation for the modules of the Prostate-MRI_Lesion_Detection, v2.0
9+
software package, which is a collection of operators which accept (T2, ADC, and High
10+
b-value DICOM images) and produce prostate organ and lesion segmentation files
11+
12+
RECIPIENT: the party that downloads the software.
13+
14+
By downloading or otherwise receiving the SOFTWARE, RECIPIENT may
15+
use and/or redistribute the SOFTWARE, with or without modification,
16+
subject to RECIPIENT’s agreement to the following terms:
17+
18+
1. THE SOFTWARE SHALL NOT BE USED IN THE TREATMENT OR DIAGNOSIS
19+
OF HUMAN SUBJECTS. RECIPIENT is responsible for
20+
compliance with all laws and regulations applicable to the use
21+
of the SOFTWARE.
22+
23+
2. THE SOFTWARE is distributed for NON-COMMERCIAL RESEARCH PURPOSES ONLY. RECIPIENT is
24+
responsible for appropriate-use compliance.
25+
26+
3. RECIPIENT agrees to acknowledge PROVIDER’s contribution and
27+
the name of the author of the SOFTWARE in all written publications
28+
containing any data or information regarding or resulting from use
29+
of the SOFTWARE.
30+
31+
4. THE SOFTWARE IS PROVIDED "AS IS" AND ANY EXPRESS OR IMPLIED
32+
WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
33+
OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT
34+
ARE DISCLAIMED. IN NO EVENT SHALL THE PROVIDER OR THE INDIVIDUAL DEVELOPERS
35+
BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
36+
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
37+
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
38+
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
39+
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
40+
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
41+
THE POSSIBILITY OF SUCH DAMAGE.
42+
43+
5. RECIPIENT agrees not to use any trademarks, service marks, trade names,
44+
logos or product names of NVIDIA, NCI or NIH to endorse or promote products derived
45+
from the SOFTWARE without specific, prior and written permission.
46+
47+
6. For sake of clarity, and not by way of limitation, RECIPIENT may add its
48+
own copyright statement to its modifications or derivative works of the SOFTWARE
49+
and may provide additional or different license terms and conditions in its
50+
sublicenses of modifications or derivative works of the SOFTWARE provided that
51+
RECIPIENT’s use, reproduction, and distribution of the SOFTWARE otherwise complies
52+
with the conditions stated in this Agreement. Whenever Recipient distributes or
53+
redistributes the SOFTWARE, a copy of this Agreement must be included with
54+
each copy of the SOFTWARE.

‎prostate-mri-lesion-seg/README.md‎

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -1,22 +1,22 @@
11
# Prostate MRI Lesion Segmentation - MONAI Deploy MAP
22

33
<p float="left">
4-
<img src="imgs/organ_seg.png" width="200" />
5-
<img src="imgs/lesion_prob.png" width="200" />
6-
<img src="imgs/lesion_mask.png" width="200" />
4+
<img src="imgs/organ_seg.png" width="300" />
5+
<img src="imgs/lesion_prob.png" width="300" />
6+
<img src="imgs/lesion_mask.png" width="300" />
77
</p>
88

99
This workflow takes T2, ADC, and HighB MRI series as input and produces several NIfTI files as output. These outputs contain organ and lesion segmentations and lesion probability maps.
1010

1111
## Software and Setup
1212

13-
In order to run this workflow and build a MAP you will need to [install MONAI Deploy App SDK](https://docs.monai.io/projects/monai-deploy-app-sdk/en/latest/getting_started/installing_app_sdk.html). The workflow is currently verified against version 0.5.1. On most systems this can be easily done with the command `pip install monai-deploy-app-sdk==0.5.1`.
13+
In order to run this workflow and build a MAP you will need to [install MONAI Deploy App SDK](https://docs.monai.io/projects/monai-deploy-app-sdk/en/latest/getting_started/installing_app_sdk.html). The workflow is currently verified against version 2.0.0. On most systems this can be easily done with the command `pip install monai-deploy-app-sdk==2.0.0`.
1414

1515
It is also recommended to have an NVIDIA GPU with at least 12 GB of memory available.
1616

1717
## Using this Repository
1818

19-
The easiest way to get started with this workflow is to run on a test image taken from the [ProstateX](https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=23691656) dataset. The `tutorial.ipynb` file walks through this process using ProstateX-0004 which can be [downloaded separately](https://drive.google.com/drive/folders/17oOgs1jIgJUxtkBRLATUdNrLSNQCLpCn?usp=drive_link) and placed in a `test-data/` directory.
19+
The easiest way to get started with this workflow is to run on a test image taken from the [ProstateX](https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=23691656) dataset. The `build_and_run.ipynb` file walks through this process using ProstateX-0004 which can be [downloaded separately](https://drive.google.com/drive/folders/1besSncSLlbeiv7UWveRJoOYQXOzu3JkU?usp=sharing) and placed in a `test-data/` directory.
2020

2121
After validating on test data, you can test this image on your own study or dataset. One of the main considerations when adapting to a new dataset will be the making sure the [DICOM Series Selector Operator](https://docs.monai.io/projects/monai-deploy-app-sdk/en/latest/modules/_autosummary/monai.deploy.operators.DICOMSeriesSelectorOperator.html#monai.deploy.operators.DICOMSeriesSelectorOperator) is configured to properly differentiate between the different naming schemes and properties of the new dataset.
2222

@@ -28,7 +28,7 @@ There are also scripts in the `scripts/` directory that test the workflow locall
2828

2929
## Models
3030

31-
The models needed to build and execute the pipeline (1 organ segmentation model, 5 lesion segmentation models) are hosted separately on [Google Drive here](https://drive.google.com/drive/folders/1wO4h5AON0MA3dxwnzl9cJlxjPxsfXcCF?usp=sharing).
31+
The models needed to build and execute the pipeline (1 organ segmentation model, 5 lesion segmentation models, 1 classification model) are hosted separately on [Google Drive here](https://drive.google.com/drive/folders/1EpjrlzEdV7CcaCYqGTIEzOapamP4Ag6M?usp=sharing).
3232

3333
Download these models and put them inside a folder named `prostate_mri_lesion_seg_app/models` alongside the rest of the application code. Pipeline creation and execution will not complete if the model file path is changed or renamed.
3434

0 commit comments

Comments
 (0)