Abstract: This project is about recognizing the object functionality in video and images. The object functionality aims to define the purpose of the object in the images, On the one hand, observing the functionality of an object (e.g., how humans interact with it) provides a strong cue for us to recognize the category of the object. On the other hand, inferring object functionality itself is an exciting and useful task. For example, one goal in robotic vision is not to tell a robot “this is a violin,” but to teach the robot how to make use of the functionality of the violin - how to play it. Further, learning object functionality also potentially facilitates other tasks in computer vision (e.g., scene understanding) or even the other fields (e.g., exploring the relationship between different objects). This provides students with machine learning methods, 3D geometry, object recognition, scene understanding, and opportunities to solve research and industrial problems.
Contact: Saeed Anwar