Abstract
In this paper, we demonstrate and compare FPGA implementations of Real AdaBoost classifiers and linear SVM classifiers for image-based human detection using histograms of oriented gradients (HOG) features, in terms of performance, hardware amount and accuracy of detection. In both architectures, a deep-pipelined stream structure and fixed-point arithmetic are employed. The evaluation results show the comparative analysis of the performance, resources, accuracy of detection between Real AdaBoost and the SVM. While FPGA resources required for Real AdaBoost designs are increased with the number of weak classifiers, the largest Real AdaBoost design with 1, 737 weak classifiers can be implemented using approximately \(60\,\%\) LUTs compared to the SVM counterpart. Although software implementation of the Real AdaBoost is much slower than the SVM due to serial evaluation of weak classifiers, the AdaBoost achieves a slightly better throughput on an FPGA, taking an advantage of parallel processing. For the detection accuracy, the AdaBoost designs are better than the SVM designs when the same number of training data is utilized. As regards the embedded use with low power requirements, the AdaBoost approach is suitable, while the linear SVM approach has a possibility to achieve a high degree of accuracy on rather large high-end FPGA systems.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Dalal, N., Triggs, B.: Histograms of oriented gradients for human detection. In: Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005, pp. 886–893 (2005)
Mizuno, K., Terachi, Y., Takagi, K., Izumi, S., Kawaguchi, H., Yoshimoto, M.: An FPGA implementation of a HOG-based object detection processor. IPSJ Trans. Syst. LSI Des. Methodol. 6, 42–51 (2013)
Cao, T.P., Elton, D., Deng, G.: Fast buffering for FPGA implementation of vision-based object recognition systems. J. Real-Time Image Process. 7(3), 173–183 (2012)
Komorkiewicz, M., Kluczewski, M., Gorgon, M.: Floating point HOG implementation for real-time multiple object detection. In: Proceedings of 22nd International Conference on Field Programmable Logic and Applications(FPL), pp. 711–714 (2012)
Hahnle, M., Saxen, F., Hisung, M., Brunsmann, U., Doll, K.: FPGA-based real-time pedestrian detection on high-resolution images. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, CVPR, pp. 629–635 (2013)
Hsiao, P.-Y., Lin, S.-Y., Huang, S.-S.: An FPGA based human detection system with embedded platform. Microelectron. Eng. 138, 42–46 (2015)
Negi, K., Dohi, K., Shibata, Y., Oguri, K.: Deep pipelined one-chip FPGA implementation of a real-time image-based human detection algorithm. In: Proceedings of International Conference on Field-Programmable Technology (FPT), pp. 1–8. IEEE (2011)
Dohi, K., Negi, K., Yuichiro, S., Kiyoshi, O.: FPGA implementation of human detection by HOG features with AdaBoost. IEICE Trans. Inf. Syst. 96(8), 1676–1684 (2013)
Matsubayashi, H., Nino, S., Aramaki, T., Shibata, Y., Oguri, K.: Retrieving3-d information with FPGA-based stream processing. In: Proceedings of the 16th International ACM/SIGDA Symposium on Field Programmable Gate Arrays, p. 261. ACM (2008)
Dohi, K., Yorita, Y., Shibata, Y., Oguri, K.: Pattern compression of FAST corner detection for efficient hardware implementation. In: Proceedings of International Conference on Field Programmable Logic and Applications (FPL), pp. 478–481. IEEE (2011)
Dohi, K., Hatanaka, Y., Negi, K., Shibata, Y., Oguri, K.: Deep-pipelined FPGA implementation of ellipse estimation for eye tracking. In: Proceedings of 22nd International Conference on Field Programmable Logic and Applications (FPL), pp. 458–463. IEEE (2012)
Yamauchi, Y., Matsushima, C., Yamashita, T., Fujiyoshi, H.: Relational HOG feature with wild-card for object detection. In: Proceedings of IEEE International Conference on ComputerVision Workshops(ICCV Workshops), pp. 1785–1792 (2011)
Fan, R.-E., Chang, K.-W., Hsieh, C.-J., Wang, X.-R., Lin, C.-J.: LIBLINEAR: a library for large linear classification. J. Mach. Learn. Res. 9, 1871–1874 (2008)
NICTA Pedestrian Dataset. https://www.nicta.com.au/category/research/computer-vision/tools/automap-datasets/
LIBLINEAR library. http://www.csie.ntu.edu.tw/~cjlin/liblinear/
Raspberry Pi Model B+ platform. https://www.raspberrypi.org/products/model-b-plus/
OpenCV 2.4.1. http://opencv.org/documentation/opencv-2-4-1.html
Nakahara, H., Sasao, T.: A deep convolutional neural network based on nested residue number system. In: Proceedings of 25th International Conference on Field Programmable Logic and Applications, FPL 2015, pp. 1–6. IEEE (2015)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this paper
Cite this paper
Oishi, M., Hayashida, Y., Fujita, R., Shibata, Y., Oguri, K. (2016). A Comparison of Machine Learning Classifiers for FPGA Implementation of HOG-Based Human Detection. In: Bonato, V., Bouganis, C., Gorgon, M. (eds) Applied Reconfigurable Computing. ARC 2016. Lecture Notes in Computer Science(), vol 9625. Springer, Cham. https://doi.org/10.1007/978-3-319-30481-6_8
Download citation
DOI: https://doi.org/10.1007/978-3-319-30481-6_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-30480-9
Online ISBN: 978-3-319-30481-6
eBook Packages: Computer ScienceComputer Science (R0)