BUFFALO, N.Y. – Computer vision is staking its claim as artificial intelligence’s hottest research field. A new set of online courses from the University at Buffalo examines how this technology is enabling computers to visually process the world.
The Computer Vision series introduces the technology behind systems that mimic complex human vision capabilities, preparing learners with the foundation necessary to design computer vision application programs from scratch. It explores the integral elements that permit vision applications such as image-editing smartphones, self-driving cars that read traffic signs and factory robots that navigate around human co-workers.
The field, say the course instructors, can be overwhelming for someone just beginning. Junsong Yuan, PhD, associate professor in UB’s Department of Computer Science and Engineering, and instructor Radhakrishna Dasari created the four-course series as a primer on computer vision and image processing.
“The first course of this specialization will give the learner an overview of the concepts and applications of computer vision, and the next three courses will cover them in detail,” Yuan says.
Courses are appropriate for individuals who possess basic programming skills and experience, and are familiar with rudimentary linear algebra, 3D coordinate systems and transformations, and simple calculus and probability.
UB’s Center for Industrial Effectiveness (TCIE), the business outreach arm for the School of Engineering and Applied Sciences, managed course production. MathWorks, developer of mathematical computing software for engineers and scientists, is partnering to provide access to MATLAB®. Learners will use the industry standard tool and its apps to implement fundamental concepts through course project work in an online lab environment.
“Students completing this specialization will be more equipped for a career in computer vision and gain valuable experience with MATLAB, an in-demand software package that’s a required skill for many jobs in this area,” says Brandon Armstrong, MathWorks senior online content developer.
The first three courses are now available on the Coursera platform, while the fourth is scheduled to launch on May 13:
- Computer Vision Basics: introduces core concepts of the computer vision field and its mission of making computers see and interpret the world as humans do.
- Image Processing, Features & Segmentation: empowers learners to develop image processing programs and leverage MATLAB functionalities to implement sophisticated image applications.
- Stereo Vision, Dense Motion & Tracking: teaches learners how to develop 3D vision applications using a stereo imaging System.
- Visual Recognition & Understanding: prepares learners to solve computer vision problems with deep learning.
Content consists of 5- to 10-minute video lesson learning sprints, demos, hands-on exercises, project work, readings and discussions. Learners gain experience writing computer vision programs through online labs using MATLAB and supporting toolboxes.
Learners may sign up for individual courses or the complete series. There is no charge to “audit” a course, which includes videos, readings, community discussion forums and the ability to view assignments. The fee to gain complete access – which includes submitting all assignments for feedback or a grade, and the opportunity to earn a verified certificate for the complete series – is $49 per month.