/breast_cancer_detection

Primary LanguagePython

breast_cancer_detection

This project uses 4 neural network structures to train the recognition of cancerous lesions, using mammographic and thermographic images as a database. As results, metrics relating to the training of each neural network are obtained.

Requirements

For this project, we recommend using Nvida's CUDNN and ANACONDA.

Instalation

You also need to run this script to set up a directory structure and install dependencies:

chmod +x setup_script.sh
./setup_script.sh

Database

After downloading the respective databases, we suggest that you extract the files into the directories of each database contained in /databases

  • Mias Mammograph | Donwload

    • After downloading the database, extract its contents into the /databasesmias_database direrectory

  • Mammotherm | Donwload

    • After downloading the database, extract its contents into the /mammotherm_database direrectory

Run Application

Once this is done, and after completing the installation of the dependencies, you can choose which database to train:

    python train_mias.py
    pyton train_mammotherm.py