Two-Dimensional Joint Bayesian Method for Face Verification


Sunghyu Han, Il-Yong Lee, Jung-Ho Ahn, Journal of Information Processing Systems Vol. 12, No. 3, pp. 381-391, Sep. 2016  

10.3745/JIPS.02.0036
Keywords: Face Verification, Joint Bayesian Method, LBP, LFW Database, Two-Dimensional Joint Bayesian Method
Fulltext:

Abstract

The Joint Bayesian (JB) method has been used in most state-of-the-art methods for face verification. However, since the publication of the original JB method in 2012, no improved verification method has been proposed. A lot of studies on face verification have been focused on extracting good features to improve the performance in the challenging Labeled Faces in the Wild (LFW) database. In this paper, we propose an improved version of the JB method, called the two-dimensional Joint Bayesian (2D-JB) method. It is very simple but effective in both the training and test phases. We separated two symmetric terms from the three terms of the JB log likelihood ratio function. Using the two terms as a two-dimensional vector, we learned a decision line to classify same and not-same cases. Our experimental results show that the proposed 2D-JB method significantly outperforms the original JB method by more than 1% in the LFW database.


Statistics
Show / Hide Statistics

Statistics (Cumulative Counts from November 1st, 2017)
Multiple requests among the same browser session are counted as one view.
If you mouse over a chart, the values of data points will be shown.




Cite this article
[APA Style]
Han, S., Lee, I., & Ahn, J. (2016). Two-Dimensional Joint Bayesian Method for Face Verification. Journal of Information Processing Systems, 12(3), 381-391. DOI: 10.3745/JIPS.02.0036.

[IEEE Style]
S. Han, I. Lee, J. Ahn, "Two-Dimensional Joint Bayesian Method for Face Verification," Journal of Information Processing Systems, vol. 12, no. 3, pp. 381-391, 2016. DOI: 10.3745/JIPS.02.0036.

[ACM Style]
Sunghyu Han, Il-Yong Lee, and Jung-Ho Ahn. 2016. Two-Dimensional Joint Bayesian Method for Face Verification. Journal of Information Processing Systems, 12, 3, (2016), 381-391. DOI: 10.3745/JIPS.02.0036.