TU Darmstadt / ULB / TUbiblio

Generalization of Fitness Exercise Recognition from Doppler Measurements by Domain-Adaption and Few-Shot Learning

Fu, Biying and Damer, Naser and Kirchbuchner, Florian and Kuijper, Arjan (2021):
Generalization of Fitness Exercise Recognition from Doppler Measurements by Domain-Adaption and Few-Shot Learning.
In: Pattern Recognition : ICPR International Workshops and Challenges, pp. 203-218,
Springer, 25th Interantional Conference on Pattern Recognition (ICPR 2021), virtual Conference, 10.-11.01.2021, ISSN 0302-9743, ISBN 978-3-030-68762-5,
DOI: 10.1007/978-3-030-68799-1_15,
[Conference or Workshop Item]

Abstract

In previous works, a mobile application was developed using an unmodified commercial smartphone to recognize whole-body exercises. The working principle was based on the ultrasound Doppler sensing with the device built-in hardware. Applying such a lab environment trained model on realistic application variations causes a significant drop in performance, and thus decimate its applicability. The reason of the reduced performance can be manifold. It could be induced by the user, environment, and device variations in realistic scenarios. Such scenarios are often more complex and diverse, which can be challenging to anticipate in the initial training data. To study and overcome this issue, this paper presents a database with controlled and uncontrolled subsets of fitness exercises. We propose two concepts to utilize small adaption data to successfully improve model generalization in an uncontrolled environment, increasing the recognition accuracy by two to six folds compared to the baseline for different users.

Item Type: Conference or Workshop Item
Erschienen: 2021
Creators: Fu, Biying and Damer, Naser and Kirchbuchner, Florian and Kuijper, Arjan
Title: Generalization of Fitness Exercise Recognition from Doppler Measurements by Domain-Adaption and Few-Shot Learning
Language: English
Abstract:

In previous works, a mobile application was developed using an unmodified commercial smartphone to recognize whole-body exercises. The working principle was based on the ultrasound Doppler sensing with the device built-in hardware. Applying such a lab environment trained model on realistic application variations causes a significant drop in performance, and thus decimate its applicability. The reason of the reduced performance can be manifold. It could be induced by the user, environment, and device variations in realistic scenarios. Such scenarios are often more complex and diverse, which can be challenging to anticipate in the initial training data. To study and overcome this issue, this paper presents a database with controlled and uncontrolled subsets of fitness exercises. We propose two concepts to utilize small adaption data to successfully improve model generalization in an uncontrolled environment, increasing the recognition accuracy by two to six folds compared to the baseline for different users.

Title of Book: Pattern Recognition : ICPR International Workshops and Challenges
Publisher: Springer
ISBN: 978-3-030-68762-5
Uncontrolled Keywords: Human activity recognition, Mobile sensors, Physical activity monitoring, Domain adaptation
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Interactive Graphics Systems
20 Department of Computer Science > Mathematical and Applied Visual Computing
Event Title: 25th Interantional Conference on Pattern Recognition (ICPR 2021)
Event Location: virtual Conference
Event Dates: 10.-11.01.2021
Date Deposited: 16 Mar 2021 08:23
DOI: 10.1007/978-3-030-68799-1_15
Additional Information:

Proceedings Part I; Part of the Lecture Notes in Computer Science book series (LNCS, volume 12661)

Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details