Parker Parrinello, “Development of New Methods to Facilitate the Detection of Abnormalities in Patient-Specific CT-Scan Images”
Mentor: Mahsa Dabagh, Biomedical Engineering
We are developing in-house code to convert CT-scan images which are in dicom form to a single two-dimensional JPEG image. This code will allow viewing all CT-scan slices (about few hundred) in one single image. Our code is written in python to convert the image, enhance the quality of image, and clear up any issues that combining the slices of the dicom files have caused. The code stacked the slices and averaged the location of each one for proper alignment, then the empty space was filled to not allow interference with the bigger circumference slices. The enhancement was done by using grayscale contrast, sharpening, and noise reduction to clean up the image. In order to test this program many brain scans were collected from different patients and was ran through the program. This will result in a clear top-down image of the brain. Allowing for viewing without the need for specialized medical software instead general photo viewing software can be used. These 2D images can then be used to easily detect and isolate any abnormalities in the vasculature. Our code will lead to using a computer to decipher scans and detect small abnormalities in it while being more accurate.