Skip to content

Latest commit

 

History

History
48 lines (36 loc) · 1.92 KB

README.md

File metadata and controls

48 lines (36 loc) · 1.92 KB

Automatic Brain Shunt Valve Recognition. Feature Analysis

This repository contains the supporting code for the manuscript:

L Giancardo, O Arevalo, A Tenreiro, R Riascos and E Bonfante. MRI Compatibility: Automatic Brain Shunt Valve Recognition using Feature Engineering and Deep Convolutional Networks. Scientific Reports volume 8, Article number: 16052 (2018). https://www.nature.com/articles/s41598-018-34164-6

In this manuscript, we describe a X-rays-based implanted device identification system for MRI safety and devise a pilot study to test the feasibility of the automatic image recognition component. We evaluate machine learning-based methods to identify different cerebrospinal fluid shunt valves from 416 skull X-rays clinical images. Our best performing method identified the valve type correctly 96% [CI 94-98%] of the times (CI: confidence intervals, precision 0.96, recall 0.96, f1-score 0.96), tested using a stratified cross-validation approach to avoid chances of overfitting.

Installation

This code has been tested with Python 3.6 and pip

git clone https://github.com/lgiancaUTH/shunt-valve-mri-compatibility.git
cd shunt-valve-mri-compatibility

Optionally, create a (conda) virtual environment and activate it

conda create -n valve-env python=3.6 

source activate valve-env

and finally, install the required dependencies

pip install -r requirements.txt

Run

Run analysis using features derived from convolutional neural networks

python runExpCNN-paper.py

Run analysis using features computed with Histogram of Oriented Gradients (HOG)

python runExpHOG-paper.py

Run analysis using features computed with Local Binary Patterns (LBP)

python runExpLBP-paper.py

Other information

This code is free to use for any non-commercial purposes, provided that the original publication is cited.

We refer to the original publication for additional infomation and acknowledgements.