Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Hand Shape Recognition Using Very Deep Convolutional Neural Networks

: Rakowski, Alexander; Wandzik, Lukasz

Postprint urn:nbn:de:0011-n-5343781 (470 KByte PDF)
MD5 Fingerprint: 0562267273de156d76653967ffdea631
© ACM This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution.
Created on: 14.2.2019

Association for Computing Machinery -ACM-:
ICCCV 2018, International Conference on Control and Computer Vision. Proceedings : Singapore, June 15 - 18, 2018
New York: ACM, 2018
ISBN: 978-1-4503-6470-6
International Conference on Control and Computer Vision (ICCCV) <2018, Singapore>
Conference Paper, Electronic Publication
Fraunhofer IPK ()

This work examines the application of modern deep convolutional neural network architectures for classification tasks in the sign language domain. Transfer learning is performed by pre-training the models on the ImageNet dataset. After fine-tuning on the ASL fingerspelling and the 1 Million Hands datasets the models outperform state-of-the-art approaches on both hand shape classification tasks. Introspection of the trained models using Saliency Maps is also performed to analyze how the networks make their decisions. Finally, their robustness is investigated by occluding selected image regions.