Abnormality Feature Extraction in the Spinal CordMRI Using K-Means Clustering
In this project author is describing concept to detect abnormalities from spinal cord by extracting features from MRI images and then calculating histogram and applying KMEANS algorithm on gray scale images. Histogram are the pixel intensity values obtained from the images and it has 0 to 255 (black and white pixels) colour ranges. If there is large number of white pixels encounter in MRI then there is chance of tumour in the image. After finding variations in histogram values we can apply KMEANS algorithm to separate black and white colours into different clusters. All white colour cluster pixels can be used to identify abnormalities in MRI image.
This project consists of 3 modules:
1) Upload MRI Images: Using this module we can upload MRI images to the application.
2) Histogram Calculation: Using this module we will extract pixels from images using python OPENCV algorithm and then calculate pixel intensity and put in an array to form histogram. If all pixels in histogram contains only black colour then no abnormal changes detected in MRI image and application will stop here only. If white colour pixels encountered inside histogram then abnormal MRI image detected and then we can apply KMEANS algorithm to detect abnormal part in MRI.
3) KMEANS Algorithm: On extracted pixels we will apply KMEANS algorithm to separate different colours in different clusters and then extract and plot all white pixels to detect abnormal patch in MRI.
About histogram and KMEANS algorithm you can read concept in paper. All images used in this project are extracted from base paper and put inside MRI folder. You can upload image from MRI folder and run application.
video file
thank you for your comment
pls call me on 8125424511