medical-imaging-matching is a model developed for BC Cancer that will return a similarity index for two mammogram images. This will help BC Cancer eliminate duplicate mammograms from their data increasing accuracy for further medical research, specifically in the use of AI as duplicates or missing information within a dataset can cause bias. Within this project we have developed various models to detect image similarity as well as developed a preprocessing script to curate datasets and split them into training and testing groups.
The members of the Medical Image Matching (1) team are as follows:
- Tech Lead: Alex Rogov
- QA Lead: Emily Medema
- Client Liaison: Marieke Gutter-Spence
- Project Manager: Niklas Tecklenburg
In order to install and run this model you need to:
-
Clone this repository.
-
Install the requirements for the model with:
pip install -r requirements_dev.txt -
Navigate to the
srcfolder. -
Run the model with
python predicting/ims/main.py