Improvement Of Facial Recognition Accuracy Using Eye-Lids Movement And Tensorflow Model
Loading...
Date
2021-05
Authors
Sukri, Syazwan Syafiqah
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
In this research study, the performance of the real-time face recognition system with machine learning, as well as the performance of each Haarcascade classifier based on accuracy and speed were investigated. The subset of machine learning called deep learning was employed in the real-time face recognition system as the deep face recognition technology has improved the state-of-the-art performance. A pre-trained model named FaceNet was used and the triplet loss technique was employed to impose a margin between every pair of faces from the same person to other faces. In other words, it minimizes the distance between the anchor and the positive from the same identity and maximizes the distance between the anchor and the negative from different identities. Furthermore, the performance of the system was further investigated by implementing the Tensorflow framework to improve the system performance by the usage of the Graphics Processing Unit (GPU). Labeled Faces in Wild (LFW) dataset was used as the benchmark to test the performance of the face recognition system. Furthermore, a preliminary experiment was conducted to evaluate the performance of Haarcascade classifiers so that the best classifier can be chosen in terms of accuracy and speed. It was found that haarcascade frontalface default exhibited the best performance compared to haarcascade frontal face alt and haarcascade frontalface alt2 with accurate number of faces detected and shortest average time taken to detect faces.
Description
Keywords
Computer science