Bezold also shared that it has been a great experience to be able to work in this extremely cutting-edge field with both his peers and Dr. Pearlstein. “The work I’ve done in the Intelligent Media Processing Laboratory under Dr. Pearlstein has helped me learn about and gain experience in the Machine Learning industry, which I wish to join upon graduation.”
Senior electrical engineering major Jake Bezold recently presented his work on generative adversarial networks at the 2019 Institute of Electrical and Electronics Engineers (IEEE) Applied Imagery Pattern Recognition (AIPR) Workshop. The event was held at the storied Cosmos Club in Washington DC. The full paper includes student authors Hussain Khajanchi ‘20, Matthew Kilcher ‘20, Alexander Benasutti ‘21, and Brian Rentsch ‘20, as well as alumnus Skyler Maxwell ‘19 and faculty advisor Dr. Larry Pearlstein.
Bezold explains that the group’s “… end goal for this specific research project is to create a rover that can autonomously wander around lawns and kill weeds without using strong herbicides, which are harmful to the environment. One of the most difficult parts of doing such a task is getting a large amount of labeled data. Therefore, the specific poster I presented was on the process of creating labeled data to achieve the aforementioned goal. Using a modified Generative Adversarial Network (GAN) technique, we were able to create hyper realistic grass which is nearly indistinguishable from real pictures of grass.”