A Study on Face Recognition using Convolution Neural Networks

• 발행기관: 한국유통과학회
• 저자명: Da yeon Son Min ho Lee
• 간행물 정보: 『한국유통과학회 학술대회 논문집』2017년 동계 국제학술대회, 253~255쪽, 전체 3쪽
• 주제분류: 경제경영 > 경제학
• 파일형태: PDF
• 발행일자: 2017.12.27

영문 초록

In recent years, development of hardware and acquisition of big data has attracted the attention of deep learning technology that learns by self learning based on data in big data, distinguishes patterns, and distinguishes objects. As we enter the information society, the problem of information security is increasing. Among them, CNN developed to imitate the human visual processing process has been widely applied to the image recognition field and shows high performance. In this paper, based on the deep - running algorithm of CNN structure, we extract facial data from github and extract specific region (eye). In general, the image recognition rate through the CNN algorithm is about 70%. This paper has improved the recognition rate to 80% through the difference of illumination and distance. It can be applied to security technology using the result of the study. As a solution to this security problem, studies using deep learning have shown good performance. Among them, in the field of image recognition, CNN algorithm is used to face recognition. However, accuracy is low according to the current situation. However, we plan to apply and develop security related technology effectively through future research.

목차

Abstract

1. Introduction

2. Related Works

3. Design & Implementation

5. Conclusion and Future Research

References

키워드

Face Detection Face Recognition Networks Neural Network