Clustering-DBSCAN
Purpose
In this project I am applying clustering techniques to detect the number of clusters in the extracted brain slices of resting state functional magnetic resonance imaging (rs-fMRI) scans.
Objectives
- To perform cluster detection in the brain slices.
Description
In this project, the program will take a patient’s dataset, performs brain slice extraction on it and then detect the number of clusters present in every extracted brain slice.
Tasks
- Extract the brain slices in every image (similar to Brain Slices Extraction).
- Once I have the brain slices images, I am applying clustering techniques to detect the number of clusters present in every slice. To extract the noticeable big enough cluster, I only am reporting the number of clusters whose pixel value is greater than 135 pixels.
Files
-
clustering.py - The clustering.py will read all the images (images those end with word “thresh”) from the given data and perform slices extraction. Once I have brain slices images, I will count number of clusters every slice contains using clustering techniques like DBSCAN.
-
test.py - This file is executed and it will call the functions in clustering.py.
-
testPatient - The test.py reads a folder named ‘testPatient’ and outputs two folders. One folder named “Slices” and another folder named "Clusters". ‘Slices’ folder will further have ‘N’ number of folders where N is number of images that ends with “thresh”. Folder ‘Clusters’ will also have N number of folders and every folder will have clusters detected images along with one ‘csv’ file which will report the number of clusters for every slice in that image folder.