This project aims to develop a simple, cheap, and accurate method for early lymphedema detection through 3-D reconstruction of the arm using Kinect depth sensors. Each sensor will take an RGB-D image of the arm, and a 3-D model will be reconstructed by piecing the individual depth images together, and the arm volume and other pieces of information such as swelling locations can be calculated from this model. This data will be immensely useful for doctors to determine whether the patient has early lymphedema or not. Furthermore, after enough expert-labeled cases, we can develop a classifier to determine the likelihood of a patient having lymphedema automatically.

I want to work on this project because the technology can make a significant difference in people’s lives, and I think 3-D reconstruction is cool. I have experience with working with the PCL library and stitching images together from my current UROP and my Computational Photography class, respectively.