Visual Computing
In this theme, you apply computing techniques for acquisition, processing, analysis and synthesis of images and 3D-digital shapes. With this, you can solve visual problems that are repetitive or require expert knowledge including medical diagnosis, autonomous vehicles/robots, and industrial inspection. Your solutions allow machines to meaningfully assist or even completely take-over such visual tasks. You will extensively study deep learning to learn visual features from huge, annotated datasets. Topics include image formation, multi-scale analysis, style transfer, deep fakes, image/video classification/detection/segmentation, point-clouds, meshes, radiance fields, etc.
Year 1 |
|||
Quarter 1 |
Quarter 2 |
Quarter 3 |
Quarter 4 |
Data management and Engineering | Software Engineering and Testing for AI Systems | Responsible Data Science and AI Engineering | Research course |
Machine and Deep Learning | Theme 1 | Theme 1 | Theme 1 |
Probabilistic AI and Reasoning | Theme 2 | Theme 2 | Theme 2 |
Credits: each course in a theme is 5EC, so each theme is 15EC.
Students choose 2 themes, each of which has 3 courses in the 2nd, 3rd and 4th quarters of the 1st year. For this theme, you will take the following courses:
Q2 - Applied Image Processing
This course provides an overview of image processing with a deeper focus on applications for visual data. This includes: image representations, colour models, linear filters, image sampling, digital photography (including HDR and tone mapping), image retargeting, image abstraction and non-photorealism, multiscale image processing (pyramids, Fourier transform), geometrical image transformations (warping, morphing), and neural image processing (concept, texture synthesis, style transfer, super-resolution, denoising, editing).
Q3 - Computer Vision
This course is on automatically understanding visual content such as images and videos by deep learning. The range of topics (although, not limited to): fundamentals in vision, visual representations, object detection, per-pixel labellings, video recognition, image similarity learning, efficiency, and self-supervision. You will build on techniques for working image data and expand them to learn how to analyse content of images and videos for classification, segmentation, or retrieval. This course aims to not just produce another modified image but rather to understand what is conveyed in the original input image and derive novel, often non-visual representations (e.g., image description, object locations, motion measurements, etc).
Q4 - 3D Visual Computing
This course builds on the knowledge from Applied Image Processing (inputs and outputs are often 2D images so the same methodologies are useful and/or can be generalized from 2D to 3D) and Computer Vision (additional non-visual information extracted from the input images can be useful) and combines it with 3D geometry and rendering. This way you will learn how to construct, render and analyse 3D representations of the world. Topics that are covered in this course are: 3D visual computing techniques and concepts such as point clouds, meshes, 3D capturing, structure from motion, shape-from-X, multi-view reconstruction, shape analysis, geometric deep learning, radiance fields, image-based rendering etc.