Face recognition via deep learning has achieved a series of breakthrough in recent years. The deeply learned features are required to be discriminative and generalized enough for identifying new unseen classes without label prediction. Therefore, this paper improved the performance of deep learning by aligning the training data and enhancing the preprocessing. Moreover, the designed model, named Lightened Convolutional Neural Network model, and the center loss layer jointly to enhance the discriminative of the designed network features. The network is trained on the self-expanding CASIA-WebFace database and tested on the Labeled Faces in the Wild (LFW) database. Experimental results show that the proposed network model brings significant improvement in the accuracy of face recognition, compared with the original CNN model.
Robust Face Recognition Based on Convolutional Neural Network / Y. Xu, H. Ma, L. Cao, H. Cao, Y. Zhai, V. Piuri, F. Scotti - In: International conference of manufacturing science and information engineering[s.l] : DEStech Pubblications Inc., 2017. - ISBN 9781605955162. - pp. 39-44 (( Intervento presentato al 2. convegno International Conference on Manufacturing Science and Information Engineering tenutosi a Guangzhou nel 2017.
|Titolo:||Robust Face Recognition Based on Convolutional Neural Network|
|Settore Scientifico Disciplinare:||Settore INF/01 - Informatica|
Settore ING-INF/05 - Sistemi di Elaborazione delle Informazioni
|Data di pubblicazione:||2017|
|Digital Object Identifier (DOI):||http://dx.doi.org/10.12783/dtcse/icmsie2017/18635|
|Tipologia:||Book Part (author)|
|Appare nelle tipologie:||03 - Contributo in volume|