ADL4CV

Welcome to the Advanced Deep Learning for Computer Vision course offered in winter semester 2024/2025!

General Course Structure

The lecture will be held in person on Wednesdays 10am-12pm. Lecture slides will be published on this website. For the meantime, we keep the lecture slides and recordings from the previous semesters on the website as well, as a reference.

In addition, the class will feature practical projects. At the end of this course, all students are expected to deliver complete projects as teams of two persons. The project presentations during the semester and at the end of the semester will be held in person. Each team/project will be aligned with a TA, who will be responsible for all project-related questions.

We will cover the class structure and planning in more detail in the first lecture at the beginning of the course.

Lecture

See lecture links below.

Lecturer: Prof. Dr. Matthias Niessner

ECTS: 8

SWS: 5

Project

In order to secure the steady progress of each project, we will schedule two presentations during the semester. The current project status and intermediate results are required in the presentation.

In addition to the regular presentations, we will also arrange a final poster presentation at the end of the course, where every team is expected to present their project with a poster as in top-tier computer vision conferences

Every team is required to submit a 4-page project report written with the CVPR paper template.

Topics:

3D Scene Understanding
3D Reconstrunction
Neural Rendering
Vision and Language
...

Presentation dates: Timeline is available in Moodle.

Teaching Assistants: Craig Lei Li, Ziya Erkoç, David Rozenberszki, and Lukas Höllein

Lecture Slides and Recordings

Slides (WS 24/25)

Week 1 - Lecture 1: I2DL Recap, DL Best Practices Recap & Visualization

Week 2 - Lecture 2: 3D Deep Learning

Week 3 - Lecture 3: Neural Fields

Week 4 - Lecture 4: NeRF

Week 5 - Lecture 5: GANs 1

Previous Slides and Recodings (WS 22/23 - WS 23/24)

Week 1 - Lecture 1: I2DL Recap, DL Best Practices Recap & Visualization - Recording
Week 2 - Lecture 2: Siamese Neural Networks & Similarity Learning - Recording
Week 3 - Lecture 3: Autoencoders, Variational Autoencoders & Style Transfer - Recording
Week 4 - Lecture 4: Representation Learning - Recording
Week 5 - Lecture 5: Sequence Models - Recording
Week 6 - Lecture 6: GANs 1 - Recording
Week 7 - Lecture 7: GANs 2 - Recording

Previous Slides and Recodings (SS 20 - SS 22)

Week 1 - Lecture 1: Visualization and Interpretability - Recording
Week 2 - Lecture 2: Siamese Neural Networks and Similarity Learning - Recording
Week 3 - Lecture 3: Autoencoders and VAE - Recording
Week 4 - Lecture 4: Deep Learning on graphs - Recording
Week 5 - Lecture 5: Generative Neural Networks - Recording
Week 6 - Lecture 6: GAN Architectures and Conditional GANs - Recording
Week 7 - Lecture 7: Videos and Autoregression - Recording
Week 8 - Lecture 8: Neural Rendering - Recording
Week 9 - Lecture 9: Deep Learning in Higher Dimensions - Recording

Prerequisites

Introduction to Deep Learning.
Strong mathematical background: linear algebra, calculus.
Previous knowledge of Python is needed. This includes scripting skills such as operating system calls.

Forum

We will use Moodle for announcements, discussions and course-related information.

Contact us

If you have any questions regarding the organization of the course, do not hesitate to contact us at: adl4cv-i28@in.tum.de. Please refrain from using the personal email addresses.

For questions on the syllabus, exercises or any other questions on the content of the lecture, we will use the forum discussion board.