Variable Tasks and Domains in Dense Prediction

Download
Full text (PDF, 2.846Mb)
Thumbnail
Open access
Author
Brüggemann, David
Date
2023
Type
    Doctoral Thesis
Rights / license
In Copyright - Non-Commercial Use Permitted
Abstract
Standard supervised learning, which dominates recognition tasks in computer vision, requires retraining models from scratch for new tasks or domains. This is inefficient and ignores related information or previous knowledge that could help learn tasks more efficiently and robustly. In this thesis, Show more
Permanent link
https://doi.org/10.3929/ethz-b-000648597
Publication status
published
External links
Search print copy at ETH Library
Contributors
Examiner: Van Gool, Luc
Examiner: Roth, Stefan
Examiner: Sattler, Torsten
Examiner: Sakaridis, Christos
Publisher
ETH Zurich
Organisational unit
03514 - Van Gool, Luc (emeritus) / Van Gool, Luc (emeritus)  
More
Show all metadata
ETH Bibliography
yes
Altmetrics