Abstract
Ethnic group classification is a well-researched problem, which has been pursued mainly during the past two decades via traditional approaches of image processing and machine learning. In this paper, we propose a method of classifying an image face into an ethnic group by applying transfer learning from a previously trained classification network for large-scale data recognition. Our proposed method yields state-of-the-art success rates of 99.02%, 99.76%, 99.2%, and 96.7%, respectively, for the four ethnic groups: African, Asian, Caucasian, and Indian.
Original language | English |
---|---|
Title of host publication | Artificial Neural Networks and Machine Learning – ICANN 2018 - 27th International Conference on Artificial Neural Networks, 2018, Proceedings |
Editors | Yannis Manolopoulos, Barbara Hammer, Vera Kurkova, Lazaros Iliadis, Ilias Maglogiannis |
Publisher | Springer Verlag |
Pages | 604-612 |
Number of pages | 9 |
ISBN (Print) | 9783030014230 |
DOIs | |
State | Published - 2018 |
Event | 27th International Conference on Artificial Neural Networks, ICANN 2018 - Rhodes, Greece Duration: 4 Oct 2018 → 7 Oct 2018 |
Publication series
Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 11141 LNCS |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Conference
Conference | 27th International Conference on Artificial Neural Networks, ICANN 2018 |
---|---|
Country/Territory | Greece |
City | Rhodes |
Period | 4/10/18 → 7/10/18 |
Bibliographical note
Publisher Copyright:© Springer Nature Switzerland AG 2018.
Funding
As previously indicated, the purpose of this research is to distinguish between the four ethnic groups: African. Asian, Caucasian, and Indian. We created our dataset by combining 10 different databases, originally proposed for the problem of face recognition, and then sorting them into the ethnic groups of interest. The databases included IMFDB [16], CNBC [20], Labeled Faces in the Wild (LFW) [7], the Essex face dataset [18], Face Tracer [10], the Yale face database [3], SCUT5000 [24], and additional collected image datasets. We also used the well-known FERET database [14,15], which contains facial images collected under the FERET program, sponsored at the time by the U.S. Department of Defense (DoD). Altogether, the collected dataset contains images of various sizes.
Funders | Funder number |
---|---|
U.S. Department of Defense | DoD |
Directorate for Social, Behavioral and Economic Sciences | 0339122 |