Type of Credit: Elective
Credit(s)
Number of Students
Deep learning has revolutionized digital image processing, pushing the limits of what one thought traditional image processing was capable of. We may think that conventional image processing techniques are washed up and no longer important. However, it may not be true. Traditional and deep-learning approaches have their own advantages and limitations. In camera systems, all the images/videos presented to you have gone through an image signal processing pipeline, which involves various image/video processing tasks, such as noise removal, contrast enhancement, gamma correction, image/video compression, image/video restoration, etc. Students will learn about these tasks and how they can be done with traditional, deep-learning, or even hybrid approaches.
能力項目說明
The course aims to teach students basics about image processing using traditional and deep-learning techniques. After the class, the students should be equipped with enough knowledge to conduct research in this field. The outline of the course includes:
| 教學週次Course Week | 彈性補充教學週次Flexible Supplemental Instruction Week | 彈性補充教學類別Flexible Supplemental Instruction Type |
|---|---|---|
|
|
Students are required to complete homework assignments, take the midterm exam, and finish a final project. They will be graded based on these three parts plus quizzes as
Homework & Labs: 30%
Mid-term exam: 30%
Final project: 40%
Quiz: 10%
References
© 2022 Richard Szeliski, The University of Washington