Analysis of optimizers on AlexNet architecture for face biometric authentication system
Nowadays, biometric authentication is more important than a password or token-based authentication. There have been many techniques suggested for biometric authentication algorithms, however, it can be observed that the Deep Learning approach is significantly more effective and secure than other met...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Proceeding Paper |
| Language: | en en |
| Published: |
IEEE
2022
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/101908/13/101908_Analysis%20of%20optimizers%20on%20AlexNet%20architecture.pdf http://irep.iium.edu.my/101908/14/101908_Analysis%20of%20optimizers%20on%20AlexNet%20architecture_SCOPUS.pdf http://irep.iium.edu.my/101908/ https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9970238 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Nowadays, biometric authentication is more important than a password or token-based authentication. There have been many techniques suggested for biometric authentication algorithms, however, it can be observed that the Deep Learning approach is significantly more effective and secure than other methods, specifically Convolutional Neural Networks (CNN) with AlexNet architecture for face recognition. However, an optimization technique is crucial in the Deep Learning models. Therefore, this paper will analyze the best optimizers for AlexNet architecture which are SGD, AdaGrad, RMSProp, AdaDelta, Adam, and AdaMax by using the proposed face dataset includes 7 celebrity classes, each with 35 images obtained from Google Images. To enhance the size of the dataset, data augmentation was employed before it was fed into the AlexNet model. The experiment shows AdaMax performs well when compared to the other
optimizers on the proposed dataset. |
|---|