kayarre / MPhys

Accurate and fast image registration using convolutional neural networks for head and neck cancer patients.

MPhys Project

---

Accurate and fast image registration using convolutional neural networks for head and neck cancer patients.

TODO

Aims

Apply DVFs from database to visualise registration
Train CNN using available DVFs as labels
Perform image registration on unseen data to produce high-quality DVFs

Weekly Progress

Week 1
Week 2
Week 3
Week 4
Week 5
Week 6
Week 7
Week 8
Week 9
Week 10
Week 11

Useful Links

General Help

Python Homepage
StackOverflow
Dicom Keyword Dictionary
Non-rigid Image Registration in Python
Slicer Python Scripting

Ebooks

How to think like a Computer Scientist
Python Machine Learning
Artificial Intelligence: A modern approach

Papers

Unsupervised End-to-end Learning for Deformable Medical Image Registration
Loss Functions for Neural Networks for Image Processing
An Unsupervised Learning Model for Deformable Medical Image Registration
3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation
HomographyNet
Multi-Temporal Remote Sensing Image Registration Using Deep Convolutional Features Github

Libraries

Keras
Medical Image Registration Library
Iterative Closest Point Algorithm
Elastix
NiftyNet
RegNet

Reports

Donal Report

Requirements

Python version: 3.6.6
bash 4.4.19, elastix 4.9, pydicom 1.2.1, SimpleITK 1.1.0, pyelastix 1.1, 3D Slicer 4.10 and SimpleElastix 1.1.