University of Groningen
Computer vision techniques for calibration, localization and recognition Lopez Antequera, Manuel
DOI:
10.33612/diss.112968625
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Lopez Antequera, M. (2020). Computer vision techniques for calibration, localization and recognition. University of Groningen. https://doi.org/10.33612/diss.112968625
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Stellingen
behorende bij het proefschrift
Computer Vision Techniques for Calibration,
Localization and Recognition
van
Manuel L ´opez Antequera
First
Learning-based methods can exploit subtle cues to predict a camera’s intrinsic parameters as well as its orientation with respect to the local gravity from a single image, outperforming traditional two-stage approaches that rely on detecting explicit geometric patterns such as lines or vanishing points.
Second
Learning-based techniques can be used to train low-dimensional representations for whole images that are discriminative with respect to the location of the cam-era, while being invariant to unrelated effects such as different illumination and weather conditions.
Third
The response of image-level representations with respect to changes in the pose of a camera can be modeled using a Gaussian Process. It can then be used as an observation model for a particle filter, enabling robust online visual localization using image-level descriptors.
Fourth
Contemporary machine learning techniques might achieve surprisingly good re-sults when used as black boxes, however, proper use of domain knowledge will make the difference between good results and excellent results.
Fifth
Don’t learn what you already know: It’s a waste of time and energy to use learning-based techniques to model a system if we can already do so effectively and efficiently with other techniques.
Sixth
Excellence rarely occurs in isolation. An association of competent individuals will produce far greater results than what could be accomplished separately.