Introduction to our open-sourced PLDC system
This open-sourced deep-learning-based model acts as an end-to-end system, input from prostate mpMRI sequences (i.e. T2, ADC, and hDWI), output to prediction results (i.e. prostate segmentation, coarse lesion detection, and malignancy estimation). The system supports multi-format inputs, including DICOM, jpeg, png, and jpg files. It is emphasized that no manual prostate segmentation or annotation is required.
1. To realize automatic PLDC with multi-cohort mpMRIs (i.e., T2, ADC, and hDWI), please download the executable software first from the
Github repository. In addition, you should also download the PLDC_software.zip from the Github repository. Unzip the zip file, and put the downloaded executable software in the folder "./PLDC_software/ ".
2. Install required packages for mpMRI-based PLDC testing.
- python=3.6.5
- Keras
- Tensorflow = 1.15
- Opencv-python
- Pydicom
- Numpy
- Pillow
- Scikit-image
- SimpleITK
3. In order to perform PLDC using your local cohort samples, you should train a domain adaptation (DA) model first (see details in the next Section " Prostate lesion assessment using your local cohort mpMRI "). Put all of your well-trained model weights in the folder "./PLDC_software/doc/weights/ ".
4. Begin to test the target mpMRIs from your local cohort using the open-sourced system. Open the executable software. Start your testing via the "Main menu" button, and then click "Start testing". The predicted results will be saved in the folder "./media/output/ ", including prostate segmentation, prostate lesion detection, and lesion malignancy results. You can download the following prostate_exe.mp4 to learn the details if necessary.