Mech DAMP Blog

CS663 - Fundamentals of Digital Image Processing

CS663 - Fundamentals of Digital Image Processing

Instructor

Prof. Ajit Rajwade

Semester

Autumn 2023-24

Course Difficulty

The course was moderate in my opinion. On a scale of 5, I would give it a 3.5. People should prefer this course in 3rd year, if not then 4th year because the professor did not allow 2nd-year undergrads.

The course was heavy only in terms of the math behind the topics (sometimes challenging but worth visualising the results) which was quite elaborate and took some time to follow up. However, the good news is that the exams are not math-heavy but rather focused more on conceptual clarity.

Time commitment needed

In terms of time commitment, you can be assured of getting a good grade if you understand the concepts well since the assignments are hands-on with the concepts you have learned so far. Almost all the questions in exams too are based on similar concepts but just a different application and I assure you that you will love writing the exams if you take interest in the course.

A lot of engaging discussion happens in classes too which should be a very good reason for you to attend classes regularly.

Grading Statistics

AA 26 AB 22 BB 22 BC 22 CC 20 CD 8 DD 2 FR 10 II 1 Total 133

These were the grading statistics in my year. As far as I know, the grading has been good and you should be able to score well (AA/AB) if you do your assignments very well and midsems/exams fairly well.

Attendance Policy

No attendance policy but the professor took attendance for his own records.

One interesting thing is that the professor starts getting a hang of people by their names pretty soon. A strange power of him somehow ;

Teaching Style

Teaching was mostly through slides and one or two lectures were delivered in online mode due to the professor not being available.

The slides were very informative and convenient to go through, learn and revise. The professor also regularly engaged in healthy in-class discussions beyond the material discussed and it was interesting to hear some new ideas. The professor is open to discussing new ideas and ways in the domain of image processing which you can display even through your project.

Moodle was regularly used for assignment submissions and accessing material.

All the course content was also simultaneously updated on Prof�s webpage as well.

Feedback on Assignments/ Tutorials/ Projects

Assignments- Interesting, comprehensive, and application-based. Each assignment will be very briefly curated and at the end you will be able to come up with your own piece of code which will solve it (eg. Image restoration, enhancement and face recognition)

Midsems/Endsems - As already mentioned earlier, midsems/endsems are more conceptual based and a good grasp over the subject will make them a cakewalk for you.

Projects - An exclusive list of very interesting projects will be provided by the professor himself. You can choose from them or you can choose your own idea and discuss with the professor. Just FYI, our project was based on Image Deblurring where we introduced a novel-method based on reverse Perona-Malik model which had not been explored earlier

Feedback on Exams (Written Evaluation)

The exams were NOT difficult NOR easy but rather involving.

Almost 50-60% of the questions are straightforward and from the slides/discussions but the other part of the paper is the decider, it is more of an open-ended answer question where you are supposed to think independently and uniquely BUT along the lines taught in class.

Future Tracks

This course is a pre-requisite for the course CS754 - Advanced Image Processing

Course Importance

This course will improve your skills in how images are processed and generated in various day-to-day applications like Image Restoration, Face Recognition etc. A wide variety of industries have been using these traditional methods which were available before the neural network era and work very efficiently.

Along with that, if you want to learn MATLAB, this course can be a good start for you since all the assignments will be based on MATLAB simulations.

Additional Details

The professor is very to the point about his marking schemes and there would be very few instances where you could convince him about a change in your corrected sheets/assignment marks unless the concern is due to some other logistics.

Also, deadlines are to be strictly followed where the professor gives an extra window of 9-10 hours after the submission deadline without penalty (idk how he defines deadline lol) and no submissions by any means are accepted after that.

Contact Details

Kandarp Solanki - 7016601717