Ear Recognition Using ResNet50

Title

Ear Recognition Using ResNet50

Creator

Singh D.; Raju G.

Description

Deep learning techniques have become increasingly common in biometrics over the last decade. However, due to a lack of large ear datasets, deep learning models in ear biometrics are limited. To address this drawback, researchers use transfer learning based on various pre-trained models. Conventional machine learning algorithms using traditional feature extraction techniques produce low recognition results for the unconstrained ear dataset AWE. In this paper, an ear recognition model based on the ResNet-50 pretrained architecture outperforms traditional methods in terms of recognition accuracy in AWE dataset. A new feature level fusion of ResNet50 and GLBP feature is also experimented to improve the recognition accuracy compared to traditional features. 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Source

Lecture Notes in Electrical Engineering, Vol-828, pp. 545-551.

Date

2022-01-01

Publisher

Springer Science and Business Media Deutschland GmbH

Subject

Deep learning; Ear recognition; GLBP; KNN; ResNet50

Coverage

Singh D., Christ (Deemed to be) University, Pune, India; Raju G., Department of Computer Science and Engineering Christ (Deemed to be) University, Bengaluru, India

Rights

Restricted Access

Relation

ISSN: 18761100; ISBN: 978-981167984-1

Format

Online

Language

English

Type

Conference paper

Identifier

https://doi.org/10.1007/978-981-16-7985-8_55

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85130910917&doi=10.1007%2f978-981-16-7985-8_55&partnerID=40&md5=7ec08110bf99d037d9b2aee57c92ece9