Welcome to the repository for 5LSM0: Neural Networks for Computer Vision, a course offered by the Department of Electrical Engineering at Eindhoven University of Technology. This course is hosted by the Video Coding & Architectures research group.
This repository contains all the assignments and supplementary materials for the course. The weekly assignments are designed as Jupyter Notebooks, providing practical, hands-on experience with the concepts discussed during lectures. These notebooks will help you gain familiarity with implementing neural networks using PyTorch in Python. For the final assignment, you will apply the knowledge gained throughout the course by coding in native Python and working with a compute cluster, providing valuable experience in real-world computational environments.
The weekly assignments are structured to guide you through foundational and advanced topics in neural networks for computer vision. These assignments are:
- Optional: They are not mandatory but serve as valuable practice to build your coding skills.
- Hands-On: Focused on applying theoretical knowledge from the lectures into real-world implementations.
The final assignment is the cornerstone of this course and accounts for 50% of your final grade. In this project, you will:
- Work on a real-world problem using the CityScapes dataset.
- Train neural networks and validate their performance against established baselines.
- Document your results and insights in a detailed report.
This final assignment requires a deeper dive into the subject, pushing you to apply the knowledge and skills gained throughout the course.
This course material is developed and maintained by the following contributors:
-
Cris H.B. Claessens
Email: c.h.b.claessens@tue.nl -
Tim J.M. Jaspers
Email: t.j.m.jaspers@tue.nl -
Francisco De Espírito Santo e Caetano
Email: f.t.de.espirito.santo.e.caetano@tue.nl -
Carolus (Koen) H.J Kusters
Email: c.h.j.kusters@tue.nl -
dr. Christiaan G.A. Viviers
Email: c.g.a.vivers@tue.nl
If you have questions or need assistance, you can always reach out to us via email. However, we strongly encourage you to post your questions in the Discussions section of this GitHub repository. This way, other students can benefit from the conversations and contribute by helping each other out.