Detection and classification of potentially cancerous areas
I was tasked with assisting radiologist by programmatically designing an algorithm to help define potential areas of concern in mammography images. I decided upon using Python and the OpenCV libraries for image recognition as they make image recognition and manipulation easy. The initial method of interpreting mammography images would either draw a circle or a box around the potentially cancerous areas. What I have done is try and refine the methods of identification to be able to take extremely accurate calculations on each of the areas of concern. Simply put, I have used contours and masks to be able to trace a perfect line around any area of concern. This project is ongoing and is being constantly updated and there are some shortcoming with this method, which will be eliminated as I continue to work on this project. To learn more about this project, please read the initial paper below.
Images derived from the program
To find out what each of these images mean you can read the detailed report in the paper outlined below
