Detection of Human Head Direction Based on Facial Normal Algorithm
http://repository.vnu.edu.vn/handle/VNU_123/10971
Many scholars worldwide have paid special efforts in searching for advance approaches to efficiently estimate human head direction which has been successfully applied in numerous applications such as human-computer interaction, teleconferencing, virtual reality, and 3D audio rendering.
However, one of the existing shortcomings in the current literature is the violation of some ideal assumptions in practice.
Hence, this paper aims at proposing a novel algorithm based on the normal of human face to recognize human head direction by optimizing a 3D face model combined with the facial normal model.
In our experiments, a computational program was also developed based on the proposed algorithm and integrated with the surveillance system to alert the driver drowsiness.
The program intakes data from either video or webcam, and then automatically identify the critical points of facial features based on the analysis of major components on the faces; and it keeps monitoring the slant angle of the head closely and makes alarming signal whenever the driver dozes off.
From our empirical experiments, we found that our proposed algorithm effectively works in real-time basis and provides highly accurate results.
Many scholars worldwide have paid special efforts in searching for advance approaches to efficiently estimate human head direction which has been successfully applied in numerous applications such as human-computer interaction, teleconferencing, virtual reality, and 3D audio rendering.
However, one of the existing shortcomings in the current literature is the violation of some ideal assumptions in practice.
Hence, this paper aims at proposing a novel algorithm based on the normal of human face to recognize human head direction by optimizing a 3D face model combined with the facial normal model.
In our experiments, a computational program was also developed based on the proposed algorithm and integrated with the surveillance system to alert the driver drowsiness.
The program intakes data from either video or webcam, and then automatically identify the critical points of facial features based on the analysis of major components on the faces; and it keeps monitoring the slant angle of the head closely and makes alarming signal whenever the driver dozes off.
From our empirical experiments, we found that our proposed algorithm effectively works in real-time basis and provides highly accurate results.
| Title: | Detection of Human Head Direction Based on Facial Normal Algorithm |
| Authors: | Lam Thanh Hien, Do Nang Toan, Tran Van Lang |
| Keywords: | Head Direction, Facial Normal Model, Novel Algorithm, Head Detection, Face Model |
| Issue Date: | 2015 |
| Publisher: | ĐHQGHN |
| Abstract: | Many scholars worldwide have paid special efforts in searching for advance approaches to efficiently estimate human head direction which has been successfully applied in numerous applications such as human-computer interaction, teleconferencing, virtual reality, and 3D audio rendering. However, one of the existing shortcomings in the current literature is the violation of some ideal assumptions in practice. Hence, this paper aims at proposing a novel algorithm based on the normal of human face to recognize human head direction by optimizing a 3D face model combined with the facial normal model. In our experiments, a computational program was also developed based on the proposed algorithm and integrated with the surveillance system to alert the driver drowsiness. The program intakes data from either video or webcam, and then automatically identify the critical points of facial features based on the analysis of major components on the faces; and it keeps monitoring the slant angle of the head closely and makes alarming signal whenever the driver dozes off. From our empirical experiments, we found that our proposed algorithm effectively works in real-time basis and provides highly accurate results. |
| URI: | http://repository.vnu.edu.vn/handle/VNU_123/10971 |
| Appears in Collections: | ITI - Papers |
Nhận xét
Đăng nhận xét