English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

With significant advances in deep learning, many computer vision applicationshave reached the inflection point. However, these deep learning models needlarge amount of labeled data for model training and optimum parameterestimation. Limited labeled data for model training results in over-fitting andimpacts their generalization performance. However, the collection andannotation of large amount of data is a very time consuming and expensiveoperation. Further, due to privacy and security concerns, the large amount oflabeled data could not be collected for certain applications such as thoseinvolving medical field. Self-training, Co-training, and Self-ensemble methodsare three types of semi-supervised learning methods that can be used to exploitunlabeled data. In this paper, we propose self-ensemble based deep learningmodel that along with limited labeled data, harness unlabeled data forimproving the generalization performance. We evaluated the proposedself-ensemble based deep-learning model for soft-biometric gender and ageclassification. Experimental evaluation on CelebA and VISOB datasets suggestgender classification accuracy of 94.46% and 81.00%, respectively, using only1000 labeled samples and remaining 199k samples as unlabeled samples for CelebAdataset and similarly,1000 labeled samples with remaining 107k samples asunlabeled samples for VISOB dataset. Comparative evaluation suggest that thereis $5.74\%$ and $8.47\%$ improvement in the accuracy of the self-ensemble modelwhen compared with supervised model trained on the entire CelebA and VISOBdataset, respectively. We also evaluated the proposed learning method forage-group prediction on Adience dataset and it outperformed the baselinesupervised deep-learning learning model with a better exact accuracy of 55.55$\pm$ 4.28 which is 3.92% more than the baseline.

Harnessing Unlabeled Data to Improve Generalization of Biometric Gender and Age Classifiers