While considering the applications of two dimensional image segmentation I thought it would be interesting to apply the procedures to medical images such as an MRI scan. These results are a part of the pipeline for tumor extraction and can be used for surgical preparations. Segmentation is equivalently a classification problem, where the labels represent tissues present in the scan. Specifically I chose to work with three dimensional brain MR images. This is a very interesting extension of techniques used in flat images. The goal is to explore some of the modern literature on the topic and attempt to use neruoimaging packages in R to achieve an accurate segmentation.
- R: For statistical computing
- mritc: Various methods for MRI tissue classification.
- imager: Image processing library.
- ggplot2: Data visualizations.
- dplyr: Data manipulation.
- gridExtra: Functions for graphics with grid layouts.
- Derek Wayne - Initial work - Portfolio
The methods used within this project are thanks to the work done by Yongyue Zhang et al. in the paper Segmentation of Brain MR Images Through a Hidden Markov Random Field Model and the Expectation-Maximization Algorith.