The goal of my project is to develop novel approaches for hand recognition. This could be used in robot gesture recognition for understanding commands from a human. Hand recognition also has many applications ranging from sign language to inferring shapes of objects. We built a vision system that can recognize 3D structure of hand poses from pictures in a controlled setting. In order to improve the performance of the system and make it more robust, we decided to understand how humans recognize hands in pictures. We are currently analyzing the psychophysics of human hand recognition by running experiments on Amazon Mechanical Turk. Once we understand the psychophysics of how humans recognize hands, we will be able to build better hand recognition systems.

I am participating in SuperUROP because this is a great opportunity to apply what I’ ve learned throughout my studies. I have taken machine learning and computer vision classes which will help me a lot in this project. I hope to learn about how computer vision is used to solve problems.