Skip to content

The Invasive Ductal Carcinoma (IDC) Detection System is an open source computer vision program created to classify IDC positive and negative samples. This project is a complete system including a locally hosted webserver / UI / API allowing you to manage your pipeline.

License

Notifications You must be signed in to change notification settings

BreastCancerAI/IDC-Detection-System

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

68 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

IDC Classifier

IDC Classifier

Introduction

The IDC Classifier is an open source computer vision program created to classify Invasive Ductal Carcinoma (IDC) positive and negative samples. The project includes a number of sub projects using different frameworks and models such as Tensorflow & Inception V3, and Caffe & CaffeNet.

The dataset used with the IDC classifer is an open dataset: Breast Histopathology Images on Kaggle by Paul Mooney.

For classification/inference the project uses the Intel® Movidius™ Neural Compute Stick, a USB stick designed to accelerate computer vision on the edge. We use the UP Squared IoT development board for our IoT device in conjuction with the Movidius.

To read a technical article about the IDC Classifier, please visit Machine Learning and Mammography on Intel AI Academy documentation.

Intel AI DevJam

The IDC classifier UI project was created as part of my Intel AI DevJam && International Conference on Machine Learning (ICML) demo in Sweden, July 2018. The goal of this demo was to intentionally trick the classifier by using very similar, but opposite class images from a small set of testing data that I believe humans may have difficulty telling apart. The project was designed to catch false negatives as a way to reduce them, providing a safety net for incorrect classifications that could mean the difference between life and death.

The GUI uses a Windows application to communicate with a facial recognition classifier and classifiers trained to detect Invasive Ductal Carcinoma (Breast cancer) in histology images. The classifiers used in the project combine the Invasive Ductal Carcinoma (IDC) Classification Using Computer Vision & IoT and TASS Movidius Facenet Classifier projects, along with some new improvements.

DISCLAIMER

The purpose of the tutorial and source code for the IDC Classifier is to help people learn how to create computer vision projects and for people interested in the medical use case evaluate if it may help them and to expand upon. Although the the program is fairly accurate in testing, this project is not meant to be an alternative for use instead of seeking professional help. I am a developer not a doctor or expert on cancer.

Bugs/Issues

Please feel free to create issues for bugs and general issues you come across whilst using this or any other BreastCancerAI repo issues: BreastCancerAI Github Issues

Contributors

Adam Milton-Barker, Intel® Software Innovator

About

The Invasive Ductal Carcinoma (IDC) Detection System is an open source computer vision program created to classify IDC positive and negative samples. This project is a complete system including a locally hosted webserver / UI / API allowing you to manage your pipeline.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published