This practical course aims at advanced students with prior knowledge of deep learning (e.g. Introduction to Deep Learning, IN2346) and multi-view geometry (e.g. Computer Vision II, IN2228). The goal of this course is to gain practical experience with state-of-the-art computer vision models and implement innovative ideas tackling open real-world challenges.

• Dr. Nikita Araslanov
• Dr. Yan Xia
• Linus Härenstam-Nielsen
• Dominik Schnaus
• Weirong Chen
• Zhenzhang Ye

• 19.01.2024: The capacity of the course is limited (max. 18 students). To ensure that the most eligible students get a spot, please send us your CV and transcripts to gsu-ss24@vision.in.tum.de by February 14 (single PDF, max. 5MB).

Course supervisors will initially offer carefully chosen project ideas and scenarios to address. These scenarios are centred around obtaining semantic and/or geometric knowledge (e.g. estimating camera/object pose, video/panoptic segmentation) from visual data (e.g. images, videos, point clouds) using deep learning networks. The problem settings can range from fully supervised to unsupervised formulations. Each project will be assigned to a group of up to three students and supported by an experienced advisor. At the end of the course, the student groups are required to present their project results in class and submit a written report.

Prerequisites: IN2346, IN2228.

gsu-ss24@vision.in.tum.de