A FACE DETECTION METHOD USING HOG AND MTCNN ALGORITHMS

Ly Van Dat; Nguyen Tien Dung

Ly Van Dat Hung Yen University of Technology and Education
Nguyen Tien Dung Hung Yen University of Technology and Education

Keywords: Face detection, Histograms of Oriented Gradients (HOG) method, Multi-task Cascaded Convolutional Neural Networks (MTCNN) method, on-board computer Jetson TX2, deep learning

Abstract

This paper researches and evaluates two methods of face detection, namely the HOG method representing the approach of using the facial features and the MTCNN method representing the use of deep learning and neural networks. To evaluate these two methods, the theory of each method was first presented. After that, the experimental model is used on the hardware platform, which is an embedded Jetson TX2 computer with an advanced GPU graphics processing chip. The face angle parameter was used to evaluate the detection level and accuracy for each method. In addition, the experimental model also evaluates the speed of face detection from the camera for each method. The results show that the average time for face detection by HOG method is 0.16 s and MTCNN method is 0.58 s. For face-to-face frames, both methods detect very well with an accuracy rate of 100 %, with face inclination angles of 30^o, 60^o, 90^o, the MTCNN method gives more accurate results. This is also consistent with published studies. Finally, the results of this study can be used as a basis for choosing face detection methods and choosing hardware devices for recognition problems.

References

N. Dalal, B. Triggs, “Histograms of Oriented Gradients for Human Detection”. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2005.

S K Eng, H Ali, A Y Cheah, Y F Chong, Facial expression recognition in JAFFE and KDEF Datasets using histogram of oriented gradients and support vector machine, IOP Conf. Series: Materials Science and Engineering, 2019, 705, 012031, doi:10.1088/1757-899X/705/1/012031.

S. Zhang, X. Zhu, Z. Lei, H. Shi, X. Wang and S. Z. Li, “Sˆ 3FD: Single Shot Scale-Invariant Face Detector,” in Computer Vision (ICCV). 2017 IEEE International Conference on, 2017.

Kaipeng Zhang, Zhanpeng Zhang, Zhifeng Li, Yu Qiao, Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks. IEEE Signal Processing Letters (SPL), 2016, vol. 23, no. 10, pp. 1499-1503.

S. Ren, K. He, R. Girshick and J. Sun, “Faster r-cnn: Towards real-time object detection with region proposal networks”. Advances in neural information processing systems, 2015.

P. Viola and M. J. Jones, “Robust real-time face detection”. International journal of computer vision, 2004, vol. 57, pp. 137-154.

V. Jain and E. Learned-Miller, “FDDB: A Benchmark for Face Detection in Unconstrained Settings”, 2010.