22955 Research Seminar: Topics in Computer Vision and Machine Learning

Credits: 3 graduate credits in Computer Science

Prerequisites: At least three graduate courses in Computer Science. Enrollment is subject to the written approval of the faculty member responsible for seminars.

Recommended: Image Processing, Prolog and Artificial Intelligence and Data Mining

This seminar aims to introduce current research topics in computer vision and in machine learning, and to set the ground for performing research.

The seminar focuses on various topics (for example, supervised learning, unsupervised learning, dimensionality reduction methods, feature selection methods; or global and local image descriptors, correspondence between images, 3D from single images). Articles and topics will be selected by the seminar’s organizers. The number of participants is limited to 20.

Requirements include attendance at 80% (at least) of the weekly sessions; meeting with one of the seminar organizers while preparing a lecture; giving a 2-hour lecture (using PPT or equivalent presentation) which describes a series of research papers; providing a written summary of the discussion and issues raised during the presentation, together with the presentation itself.