50p
Computer Vision: Towards Building Computers That See
On October 2nd, Brentwood will welcome a member of UVic's Speakers Bureau. At 7:30 pm in the Lecture Theatre, Dr. Alexandra Branzan Albu will give a presentation on artificial intelligence.
Presentation Summary
Computer vision covers some of the most exciting topics in artificial intelligence. Since the field’s beginnings in the 1960’s, computer vision technologies have made significant progress. Such technologies are in commercial use now for face recognition, medical diagnostics, and a variety of industrial applications. We will soon be able to have smart cars equipped with pedestrian detection systems.
To build a computer vision system, all we need is a camera linked to (or embedded in) a computer. The challenge is not in the hardware set-up but in the software that we have to develop in order to make our computer ‘see’.
The talk will first explain the main principles that we need to use if we want to equip our computers with ‘basic vision skills’, as well as how computer vision is different from human or animal vision. We will focus next on a particular field of computer vision applications which deals with human motion and human activity recognition. The computer-based understanding of how humans move is very important for the security of our airports, parking lots etc. but also for interacting with the computer in new ways, for instance by replacing the mouse with hand gestures or eye-blinks.
The image shown is an example of finding the right motion representation (here, the motion history image, in center) for the purpose of our task. Please click on the image to enlarge it. Alexandra Branzan Albu is an Assistant Professor with the Department of Electrical and Computer Engineering at the University of Victoria. Her research interests include computer vision-based human motion analysis and medical imaging; more information can be found at http://www.ece.uvic.ca/~aalbu. Dr. Branzan Albu is a Professional Engineer affiliated to the Province of British Columbia Association of Professional Engineers. She is also Chair of the Technical Program Committee of the IEEE Victoria Section.
Presentation Summary
Computer vision covers some of the most exciting topics in artificial intelligence. Since the field’s beginnings in the 1960’s, computer vision technologies have made significant progress. Such technologies are in commercial use now for face recognition, medical diagnostics, and a variety of industrial applications. We will soon be able to have smart cars equipped with pedestrian detection systems.
To build a computer vision system, all we need is a camera linked to (or embedded in) a computer. The challenge is not in the hardware set-up but in the software that we have to develop in order to make our computer ‘see’.
The talk will first explain the main principles that we need to use if we want to equip our computers with ‘basic vision skills’, as well as how computer vision is different from human or animal vision. We will focus next on a particular field of computer vision applications which deals with human motion and human activity recognition. The computer-based understanding of how humans move is very important for the security of our airports, parking lots etc. but also for interacting with the computer in new ways, for instance by replacing the mouse with hand gestures or eye-blinks.
The image shown is an example of finding the right motion representation (here, the motion history image, in center) for the purpose of our task. Please click on the image to enlarge it. Alexandra Branzan Albu is an Assistant Professor with the Department of Electrical and Computer Engineering at the University of Victoria. Her research interests include computer vision-based human motion analysis and medical imaging; more information can be found at http://www.ece.uvic.ca/~aalbu. Dr. Branzan Albu is a Professional Engineer affiliated to the Province of British Columbia Association of Professional Engineers. She is also Chair of the Technical Program Committee of the IEEE Victoria Section.