Abstract
Despite face recognition and verification have achieved great success in recent years, these achievements are experimental results on fixed data sets. Implementing these outstanding technologies in the field of undeveloped data sets presents serious challenges. We adopt three state-of-the-art pre-trained models on an entire new dataset University Test System Database (UTSD), however the results are far from satisfactory. Therefore, two methods are adopted to solve this problem. The first way is data augmentation including horizontal flipping, cropping and RGB channels transform, which can solve the imbalance of label pairs. The second way is the combination of Manhattan Distance and Euclidean Distance, we call it Dual Minkowski Loss (DML). Through the implementation of photo augmentation and innovative method on UTSD, the accuracy of face verification has been significantly improved, achieving the best 99.3%.
Highlights
Nowadays, a large variety of photos, videos and text scripts were applied to deep learning
The large scale of datasets contributes a lot to the improvement of recognition accuracy, such as Labeled Faces in the Wild (LFW), YouTube Faces (YTF), CASIA-WebFace, and CAS-PEAL et al [6]
As a result we improve the method through the combination of Manhattan Distance and Euclidean Distance, which is called Dual Minkowski Loss (DML)
Summary
A large variety of photos, videos and text scripts were applied to deep learning. The large scale of datasets contributes a lot to the improvement of recognition accuracy, such as Labeled Faces in the Wild (LFW), YouTube Faces (YTF), CASIA-WebFace, and CAS-PEAL et al [6]. The recent state-of-the-art face recognition models such as DeepFace have achieved an accuracy of 97.35% on LFW dataset, the later published technique FaceNet refreshed the latest record and pushed the precision to the highest 99.63%. Due to the existence of an intermediate bottleneck layer, the operation speed and accuracy of the convolution neural network have been greatly affected. FaceNet model is transferred to our new database (UTSD) for learning, three pre-trained models are adopted as initial input. Comparing the performances on the novel database, several improved measurements are adopted for the effect of the model and accuracy of prediction
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
