Data-Driven Methods for Event Cameras (Ph.D. defense of Mathias Gehrig)

Standard cameras are the de facto standard in computer vision applications but face challenges in high-speed and dynamic scenarios. On the other hand, event cameras overcome these challenges by capturing pixel-level brightness changes asynchronously. This leads to high-speed capture with minimal data redundancy and superior dynamic range, reducing issues like motion blur. However, event cameras require specialized algorithms due to their unique data stream. Event cameras have been effective in multiple domains, but the lack of large-scale datasets for event-based vision hampers progress. This Ph.D. dissertation introduces and presents a new multi-sensor dataset with real-world data and algorithms for event-based vision tasks, leveraging the high-temporal resolution of event data for robust optical flow prediction and object detection.
Reference:
Mathias Gehrig
Data-Driven Methods for Event CamerasGoogle Scholar: scholar.google...
Our research page on event-based vision: rpg.ifi.uzh.ch/...
For event-camera datasets, see here:
1. dsec.ifi.uzh.ch/
2. rpg.ifi.uzh.ch/...
3. github.com/uzh...
For an event camera simulator: rpg.ifi.uzh.ch/...
For a survey paper on event cameras, see here:
rpg.ifi.uzh.ch...
Other resources on event cameras (publications, software, drivers, where to buy, etc.):
github.com/uzh...
Affiliation:
Mathias Gehrig is with the Robotics and Perception Group, Dept. of Informatics, University of Zurich, and Dept. of Neuroinformatics, University of Zurich and ETH Zurich, Switzerland rpg.ifi.uzh.ch/

[@eason_longleilei]

Congratulations to Gehrig.👍👍

[@jetthomaspham]

This is amazing! I would love to implement this technique myself to make more robust autonomous drone navigation

[@oscbit]

Beeindruckend!