Differences and Features Analysis of Face Recognition Using Genetic Algorithms

• Main Authors: WANG, CHING-WEI, 王敬爲
• Other Authors: CHEN, JONG-CHEN
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/nk76wu

碩士 === 國立雲林科技大學 === 資訊管理系 === 107 === This study proposes an artificial intelligence as a standard procedure method, and hopes to find out the key attributes of face image vectors by genetic algorithm and support vector machine to distinguish different user categories and discharge strangers outside the white list. , the situation of reducing the false positive rate occurs. This study uses the ResNet convolutional neural network architecture to extract the facial image 128 attribute as the main analysis data of this study, and based on these image representative points thrown into the gene algorithm and support vector machine training to find out the image of each user's face. Important attributes to improve the identification of errors. The data set used in this study consisted of 859 individuals and contained 62,244 facial images. In the experimental results, the gene algorithm method is the best for identifying whitelist users, and the support vector machine works best for the stranger. The experiment finally tries to match the gene algorithm with the support vector machine to identify the stranger result than the single. It is better to use a support vector machine.