Pucadopr / CheXNet-Model-Compression

Official Implementation of "CheXNet Model Compression for Pneumonia detection on low edged device"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Melanoma-Prediction-App

This app is a melanoma detection application/model trainer. Dataset used for training is the ISIC Archive provided by ISIC (International Skin Imaging Collaboration).

Model Performance

Model was trained using the ISIS Archive dataset but due to limitations from the fact that dataset contained mostly white people with the condition, extra data of a thousand black people with melanoma was sourced and added to the dataset for training to ensure model is more robust.

Requirements

  • Python 3.5 or newer.
  • Other requirements can be installed using the requirements.txt.

Setup

Clone the project:

git clone https://github.com/Pucadopr/Melanoma-Prediction-Model.git

Enter project directory:

cd Melanoma-Prediction-Model

Downloading the dataset:

tar -xvf faster_rcnn_inception_v2_coco_2018_01_28.tar.gz

Install required modules:

pip3 install -r requirements.txt

Train model:

python3 main.py

About

Official Implementation of "CheXNet Model Compression for Pneumonia detection on low edged device"


Languages

Language:Python 100.0%