CSE291 Machine Perception / Spring 2008

Lecture times: Tue,Thu 2:00pm - 3:30pm. Center Hall room 224A.

Labs: Several Labs in JSOE will be used for projects in this course. Students should expect to spend at least 10 hours per week on the project, outside of lecture time.

This course is a study and development of software and hardware for machine perception. The focus of the course will be on the interactive display on the fourth floor of the CSE building.

The Automatic Cameraman was an outcome of the class last year and will be a starting point for this quarter.

The course is project based. Groups of 3-4 students will work together on a project that will become part of the interactive display.

Foci of the course will include:

• Using FPGAs to perform video analysis in real time.
• Motion detection.
• face/torso detection and tracking.
• Gesture recognition.
• Audio source localization and tracking.
• speech reconstruction and speech recognition.
• Audio Scene analysis.
• Your idea here ...

Work outside lectures will require at least 10 hours per week. Lecturer will meet with each project group for one hour each week.

Prerquisites: Students taking this course should be highly motivated and self driven. Background in machine learning and digital signal processing is a plus but not a requirements. Highly motivated undergraduates are also invited. All students interested in the course should contact Professor Yoav Freund: yfreund@ucsd.edu

-- YoavFreund - 07 Mar 2007