Publication

Fake data - the future of advanced computer vision systems

Corcoran, Peter
Thumbnail Image
Files
Thumbnail Image
C3I_FakeData04rrrR.pdf
Adobe PDF, 2.95 MB
Identifiers
http://hdl.handle.net/10379/15508
 https://doi.org/10.13025/18649
Publication Date
2018-10-22
Keywords
Synthetic Data, Data Augmentation, Computer Vision, Deep Neural Networks, Artificial Intelligence, Smart City, Autonomous Driving
Type
Conference Paper
Downloads
Show all metadata
Citation
Corcoran, Peter. (2018). Fake data - the future of advanced computer vision systems. Plenary paper presented at the 2018 Duke Kunshan University International Innovation Symposium on Smart Camera Technologies, Duke Kunshan University Campus, Shanghai, China, 22-23 November, DOI: 10.13025/qbe4-qr97
Abstract
Recent research shows that Data Augmentation techniques and Synthetic Data can improve the accuracy and reduce the susceptibility of Deep Neural Networks to Adversarial Attacks. In this presentation we consider some of the new tools that are available to build advanced virtual models that can be used to render large 2D training datasets suitable for training tomorrow's advanced Computer Vision systems for deployment in consumer and smart-city use cases.
Funder
Publisher
NUI Galway
Publisher DOI
Rights
Attribution-NonCommercial-NoDerivs 3.0 Ireland
cbn
Collections
Electrical and Electronic Engineering (Conference Papers)