期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2018
卷号：96
期号：23
出版社：Journal of Theoretical and Applied
摘要：Face recognition is a technology that has been widely applied for various purposes. The example is face recognition for the creation of identity cards, driving license or security system of a place that can recognize the face of someone. In security system, surveillance using Closed Circuit Television (CCTV) camera generally use infrared camera which can capture images even in dark conditions. Generally, the image quality from CCTV camera is poor. This can be caused by a poor image retrieval process, such as inadequate capture distance and minimal lighting conditions so the result of image is still not optimal. One solution that can be done to overcome the lack of image quality by using the method of image enhancement. In this research, used two retinex method that is Single Scale Retinex and Multi Scale Retinex with Principal Component Analysis face recognition method. Testing performed compared face recognition without retinex and face recognition with retinex (SSR and MSR). Testing with the same light training data conditions has the most effective conditions when the light is very dark.
关键词：Image; Face Recognition; Infrared; Image Enhancement; Retinex
