VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.

263-5909-00L 2 Credits MSC , WBZ D-INFK

Egocentric Computer Vision: Perception and Understanding from the First-Person View

Lecturers & Examiners: Prof. Dr. Christian Holz

The deadline for deregistering expires at the end of the second week of the semester. Students who are still registered after that date, but do not attend the seminar, will officially fail the seminar.

VVZ CR n/a

Last Updated: 2026-06-03 00:14:10

Abstract

This seminar explores the rapidly evolving field of egocentric perception, focusing on how first-person vision systems and optional on-body sensors enhance human-world understanding.

Objective

The objective of this seminar is to provide students with an in-depth understanding of current research in egocentric perception, emphasizing its applications in computer vision and robotics. By analyzing and presenting cutting-edge publications, students will improve their ability to critically assess complex methodologies, synthesize findings, and articulate scientific contributions. The seminar also aims to foster collaborative discussions, helping participants refine their communication and analytical skills in a peer-supported environment.

Content

The seminar will delve into topics such as human-object interaction, scene understanding, activity recognition, and multimodal data fusion from egocentric perspectives. Participants will investigate how first-person vision systems capture the nuances of human behavior and interaction with the environment, with implications for robotics, augmented reality, and AI. Each session will feature a detailed presentation of a selected paper, followed by a group discussion led by the presenter, encouraging a comprehensive exploration of the topic.

Resources

Literature

24 papers will be provided by the lecturer and distributed in the first seminar on a first-come, first-served basis according to participants' preferences. The lecturer will also give a brief run-down across all 24 papers in a fast-forward style, covering each paper in a single-minute presentation, and outline the difficulties of each project. The schedule is fixed throughout the term with easier papers being presented earlier and more comprehensive papers presented later in the term.

Learning Materials (Links)

Main link
Information

General Information

Language: English
Levels: MSC , WBZ
Frequency: Yearly recurring

Examination

Type: graded semester performance

Students individually read one full-paper publication, working through its content in detail and covering background literature if necessary.Students then present the approach, methodology, research question and implementation as well as the evaluation and discussion in a 20–25 min talk in front of the others. Each presenter will then lead a short discussion about the paper, which is also guided by questions posed to the audience.

Registration & Places

Max Places: 24

Course Components

Type	Title	Time & Place	Hours
seminar	Egocentric Computer Vision: Perception and Understanding from the First-Person View	Tue 16:15-18:00 (STD G 1)	2 h weekly