Cost-Efficient Adaboost-based Face Detection with FPGA Hardware Accelerator

In this paper, a cost-efficient Adaboost-based hardware accelerator design is developed for face detection. For the realization, a lot of operations and fast calculations of the rectangular Haar features of partial image regions are required. To speed up the processing time for integral image and face detection functions, the hardware accelerator implementation is an important issue for real-time application. In experimental results, by Xilinx XC7a200t device at 205MHz operational frequency, the applied integral image accelerator performs up to 1161 frames/sec with the image size of 240×360 pixels. When the input ROI sizes are 80×80 and 64×64 pixels, the proposed Adaboost-based classifier operates up to 574 and 821 face ROIs/sec, respectively.

[1]  Devansh Saxena,et al.  An efficient and cost effective FPGA based implementation of the Viola-Jones face detection algorithm , 2017 .

[2]  Rainer Lienhart,et al.  An extended set of Haar-like features for rapid object detection , 2002, Proceedings. International Conference on Image Processing.

[3]  Leibo Liu,et al.  An AdaBoost-Based Face Detection System Using Parallel Configurable Architecture With Optimized Computation , 2017, IEEE Systems Journal.

[4]  Leibo Liu,et al.  A Fast Integral Image Computing Hardware Architecture With High Power and Area Efficiency , 2015, IEEE Transactions on Circuits and Systems II: Express Briefs.

[5]  Paul A. Viola,et al.  Rapid object detection using a boosted cascade of simple features , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.