Combining Learning and Hand-crafted Based Features for Face Recognition in still Images

被引：0

作者：

Sawarbandhe, Mahesh D. ^{[1
]}

Satpute, V. R. ^{[1
]}

Cheggoju, Naveen ^{[1
]}

Sinha, Saugata ^{[1
]}

机构：

[1] VNIT Nagpur, Nagpur, Maharashtra, India

来源：

PROCEEDINGS OF THE 2019 IEEE REGION 10 CONFERENCE (TENCON 2019): TECHNOLOGY, KNOWLEDGE, AND SOCIETY | 2019年

关键词：

Context-Aware Local Binary Feature Learning; Dual-Cross Patterns; Principle Component Analysis; compact binary face descriptor FERET; PATTERNS; REPRESENTATION; CLASSIFICATION; EIGENFACES; HISTOGRAM; MODEL;

D O I：

10.1109/tencon.2019.8929245

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

This paper presents a combined face recognition method based on Context-Aware Local Binary Feature Learning (CA-LBFL) which is learned based Descriptor and Dual-Cross Patterns (DCP) which is hand-crafted Descriptors. Both features are extracted and concatenated into a single enhanced feature vector which can be used as a face descriptor. In unconstrained face images the intra personal variations, pose and illumination changes are more, hence the combination of features can effectively solve this problem. The experimental results illustrate that this method has better recognition rate and more robust than single individual methods on the FERET database.

引用

页码：2502 / 2506

页数：5

共 23 条

[1] Face description with local binary patterns:: Application to face recognition [J].