Home     Schedule     FAQ
Q: Can we audit this module?
A: The lecture notes will be made available in this website. So, you can download the lecture notes and read them yourself.
However, reading the lecture notes alone is going to help you much. You need to put what you learn into practice in the form of assignments and project. So, if you are keen to learn, you might as well register for this module.
Q: Can I take this module while doing internship?
A: You need to check with your internship supervisor and the internship administrative committee about this.
In general, I have no objection about this.
Q: What does it mean by high-level, structured problems?
A: It means problems that encompass conceptual structures such as spatial structure and temporal structure. Spatial structures typically capture shape. It can also capture general relationships between spatial points such that the relationship between 3D scene points and 2D image points in camera projection, or the relationships in high-dimensional space. Temporal structures capture relationship in time.
Q: Does the module mainly focus on computer vision area?
A: No. This module does not cover the details of computer vision. Instead, it covers the foundational concept of media computing common to computer vision, computer graphics, time series, music, etc. For ease of illustration, we will focus primarily on images, 3D models, temporal sequences of images and 3D models and music.
Q: Does this module goes through deep neural networks?
A: No. There are two specific modules on neural networks and deep learning. There is no point duplicating the materials in this module.
Nevertheless, this module will relate to the basic concepts of neural networks.
Q: Will this module touch on video encoders/codecs (av1, vp9 etc)?
A: No. This course focuses on the high-level concepts. It does not cover low-level processing at the levels of signal and features. So, it does not cover signal processing, codec, streaming, feature extraction, etc. It also does not cover classification, neural networks and deep learning, which have separate modules for them.
Learning Objectives
Q: Do we get to practice what we learn in class?
A: Yes. You will get to practice applying the knowledge learned. This practice comes in two forms: assignments and team project. In each assignment, you practice one aspect of problem solving. In team project, you practice all aspects of problem solving. Programming is optional for team project.
Q: Why is programming optional?
A: Past experience of teaching this module shows that students are already well-versed with programming. What they lack, instead, is problem solving skills, in particular problem formulation and algorithm design. Therefore, the practical part of this module focuses on problem formulation and algorithm design for solving realistic media problems.
Q: For the assignments, is programming required?
A: No, because all of you are already well-versed with programming.
Q: Are there programming exercises that I can try for my own learning benefit?
A: Yes, you can find them in the website of CS5240 in LumiNUS.
Q: For the exercise and homework, will the answer be explained, or only text answers will be released?
A: Homework and exercises contain straightforward practice questions that sharpen your math skills. The answers are already given in the questions. If you have difficulty working out the answers, you may email me for help.
Q: Will the answers for assignments be explained, or only text answers are released?
A: Assignments 1 to 3 are regular assignments that have model answers. So, their answers will be released. Moreover, feedback to Assignmet 1 and 2 will be discussed to help the students better master the skill of problem formulation. Assignment 3's answer is straightforward and does not require discussion. Nevertheless, if you don't understand, you can always email me.
Assignments 4 to 6 are open questions that do not have model answers.
Q: For the team project, is demo required?
A: No, because programming is optional for this module.
Q: What is the objective of the team project?
A: In team project, you practice problem formulation and algorithm design; programming is optional.
For problem formulation, you practice describing what are required and why they are required.
For algorithm design, you practice describing how the algorithm works and why it works.
If you wish, you can implement a part of your algorithm to verify your ideas, but this is optional.
Q: What are the types of projects? Can we propose our own project?
A: The project topics must match the scope of this module. Click here for example project topics.
Yes, you are encouraged to propose your own project. In fact, many project topics in the past years were proposed by the students.
Q: How is the team project evaluated?
A: The team project is evaluated according to how well you apply the skills of problem formulation and algorithm design.
It is not dependent on how close is your algorithm to being optimal.
Q: Can I do a project on a topic with known algorithms?
A: You need to explain how your algorithm works and why it satisfies your version of the problem as defined in your problem definition.
If you cannot explain how and why your algorithm works, then you get 0 mark for algorithm design.
If you can explain how and why your algorithm works, then you get high marks, even if your algorithm is not optimal.
In summary, the project is evaluated according to how well you apply the skills of problem formulation and algorithm design,
instead of how good is your algorithm.
Q: Can you elaborate on the 10% participation?
A: I will meet each project team at least two times to discuss your project progress and to give you pointers. Your participation at these project meetings will contribute to the 10% participation. In addition, during final project presentation, you will be given opportunities to raise good questions regarding other teams' presentations. This will also count towards your 10% participation.
Q: Do we need to buy any textbooks?
A: There is no textbook available because the course materials are distilled from our own research experience and relevant papers.
The good news is that I'm writing a textbook. You can download a draft of the textbook from LumiNUS.
Q: Will we need any high computing resources for this module?
A: The high computing resources that you need for this module is you brain, or rather your thinking. A lot of emphasis is placed on understanding, analysis, reasoning and justifcation that are important for problem solving.
26 May 2021