Biology Project Abstract
3D RECONSTRUCTION OF TISSUE STRUCTURE FROM 2D IMAGES OF TISSUE SLICES
Presenter:
Rishi Bhayana, Illinois Mathematics and Science Academy, 1500 West Sullivan Road, Aurora, IL, 60506; rbhayana@imsa.edu
Mentor:
Dr. Ali Shokoufandeh, Drexel University, 3141 Chestnut Street, Philadelphia, PA, 19104; 215-895-2671; ashokouf@cs.drexel.edu
Abstract:
Recent developments in medical imaging technologies have allowed scientists to obtain 2D images of tissues in the human body in order to help in diagnosing cancer during its early formation. However, current technology only allows for images of very thin 2D slices of tissue to be taken. This results in the loss of much necessary data of the 3D tissue structure including the distribution of cells, frontier structure, and the progression of cancer cells. The purpose of this project is to create a 3D reconstruction of tissues given many 2D slices in order to retain this information. There are a variety of methods that can be used to assemble the 3D structure, but the most effective are clustering and convoluting. The process of clustering uses matrices in order to build the framework of the image in order for the computer to recognize where the data and main part of the image lies. Next, convolution, also called edge detection, is used to find the optimal place to merge the images in order to create a 3D structure. After this process is complete, pattern matching is used in order to help find similarities and differences in the framework of the images between healthy and unhealthy patients over time. Therefore, the more information that this program is provided, the more effective this program becomes in recognizing cancer during its early formations. The overall goal of this project is to create a means for which the 3D structure of tissue can be obtained from a given set of 2D images in order to be manipulated to help in the diagnosing of cancer.