| Category | Difficulty |
|---|---|
| HW | 3 |
| Exams | 4 |
This course builds on the topics of signal processing and applies them to Images and Videos which are 2D and 3D signals respectively. Throughout the course, you will learn various techniques to solve problems like denoising, inpainting, deblurring and demosaicing. Topics covered are DCT, STFT, Wavelets, Radon Transforms, Gabor Filters, OMP, IHT, Sparsity and Compression techniques and near the end of the course reviewing landmark research papers in the field, followed by a brief introduction to Computation Photography. 18-290 and 18-491 are good prerequisites to have for this course. The course is not difficult and acts as a good foundation course for other Computer Vision related courses, such as 16-720 Computer Vision and 15-862 Computational Photography.
- HW: HW consists of both derivations and programming part. Not too hard -- if you get stuck you can ask during OH or on Piazza.
- Exams: Moderately difficult so prepare well and practice assignment problems.
- Attend and understand lectures. Lecture recordings are available which is good for review.
- Start HW early
- Attend Recitations and Office Hours (OH), especially OH conducted by the professor
- Refer to Supplementary materials provided by the professor
- Follow discussions on Piazza
- You will have an assignment every week with no late days provided. Only 1 hour late submission is allowed, so plan to complete the assignment accordingly
- There is no relation between the difficulty level of the midterm and final exam, so you can't rely on your midterm performance to assess the level of preparation required for final exams. Prepare well to ace the exams.