Skip to main content

SC4003 Computer Vision

Course Summary​

This course aims to instill knowledge about how computers can see the world by transforming the raw sensor input, as well as how various filters work to affect the resultant image after post processing. The course gives very low level introductions to the various topics and does not feature more advanced recent developments such as CNNs.

note

As of writing, the school has said it is looking to refresh the content of this module so do look out for any changes.

The content that will be covered is:

  1. Imaging and Photometry
  2. Image Enhancement (Spatial and Frequency)
  3. Edge Detection
  4. Region Processing
  5. Imaging Geometry
  6. Stereo Vision
  7. Object Recognition
  8. Image Synthesis

Workload​

The workload of this course is relatively straightforward with lectures taking up 2 hours and tutorials taking up 1. However, the profs might not follow this timeline strictly and do use the slots interchangably to cover content for the tutorial. There are also Lab submissions that are to be done in MATLAB which gives you practice for the content convered in lectures such as image filtering. The labs are guided, require no prior knowledge of MATLAB code and you will generally be able to do it with some patience.

Projects​

There is an individual assignment/project which also includes tasks that have been covered in lectures although this is not as guided as the labs. You can still use the labs as a guide to help you finish the project.

Things to take note of​

The content can be pretty dry for the most part as it is just a lot of theory and some equations. Being able to visualise the concepts does help. It also does not include modern day computer vision topics like entity detection and recognition using deep learning, so you might want to wait until the content is updated before considering this module.

The profs are not the most easy to understand which makes studying the content that much harder as the slides just have the key bullet points.

Conclusion​

Take this module if you're interested in basic computer vision processing but if you want more advanced topics, hold off until a course refresh.