Pattern Classification Using Quantized Neural Networks for FPGA-Based Low-Power IoT Devices

被引:5
作者
Biswal, Manas Ranjan [1 ]
Delwar, Tahesin Samira [1 ]
Siddique, Abrar [1 ,2 ]
Behera, Prangyadarsini [1 ]
Choi, Yeji [1 ]
Ryu, Jee-Youl [1 ]
机构
[1] Pukyong Natl Univ, Dept Intelligent Robot Engn, Busan 48513, South Korea
[2] Univ Waterloo, Dept Elect & Comp Engn, Waterloo, ON N2L 3G1, Canada
基金
新加坡国家研究基金会;
关键词
internet of things (IoT); convolutional neural network (CNN); quantized neural network (QNN); binary neural network (BNN); FPGA; pattern recognition; OBJECT DETECTION;
D O I
10.3390/s22228694
中图分类号
O65 [分析化学];
学科分类号
070302 ; 081704 ;
摘要
With the recent growth of the Internet of Things (IoT) and the demand for faster computation, quantized neural networks (QNNs) or QNN-enabled IoT can offer better performance than conventional convolution neural networks (CNNs). With the aim of reducing memory access costs and increasing the computation efficiency, QNN-enabled devices are expected to transform numerous industrial applications with lower processing latency and power consumption. Another form of QNN is the binarized neural network (BNN), which has 2 bits of quantized levels. In this paper, CNN-, QNN-, and BNN-based pattern recognition techniques are implemented and analyzed on an FPGA. The FPGA hardware acts as an IoT device due to connectivity with the cloud, and QNN and BNN are considered to offer better performance in terms of low power and low resource use on hardware platforms. The CNN and QNN implementation and their comparative analysis are analyzed based on their accuracy, weight bit error, RoC curve, and execution speed. The paper also discusses various approaches that can be deployed for optimizing various CNN and QNN models with additionally available tools. The work is performed on the Xilinx Zynq 7020 series Pynq Z2 board, which serves as our FPGA-based low-power IoT device. The MNIST and CIFAR-10 databases are considered for simulation and experimentation. The work shows that the accuracy is 95.5% and 79.22% for the MNIST and CIFAR-10 databases, respectively, for full precision (32-bit), and the execution time is 5.8 ms and 18 ms for the MNIST and CIFAR-10 databases, respectively, for full precision (32-bit).
引用
收藏
页数:14
相关论文
共 31 条
[1]   Automatic detection and classification of radiolucent lesions in the mandible on panoramic radiographs using a deep learning object detection technique [J].
Ariji, Yoshiko ;
Yanashita, Yudai ;
Kutsuna, Syota ;
Muramatsu, Chisako ;
Fukuda, Motoki ;
Kise, Yoshitaka ;
Nozawa, Michihito ;
Kuwada, Chiaki ;
Fujita, Hiroshi ;
Katsumata, Akitoshi ;
Ariji, Eiichiro .
ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY ORAL RADIOLOGY, 2019, 128 (04) :424-430
[2]  
Austin J., 1998, ICANN 98. Proceedings of the 8th International Conference on Artificial Neural Networks, P469
[3]  
Bartan B., 2021, P INT C MACH LEARN, P694
[4]  
Belabed T., 2021, 2021 INT C INNOVATIO, P1, DOI DOI 10.1109/INISTA52262.2021
[5]   A Survey and Taxonomy of FPGA-based Deep Learning Accelerators [J].
Blaiech, Ahmed Ghazi ;
Ben Khalifa, Khaled ;
Valderrama, Carlos ;
Fernandes, Marcelo A. C. ;
Bedoui, Mohamed Hedi .
JOURNAL OF SYSTEMS ARCHITECTURE, 2019, 98 :331-345
[6]   A Learning Framework for n-Bit Quantized Neural Networks Toward FPGAs [J].
Chen, Jun ;
Liu, Liang ;
Liu, Yong ;
Zeng, Xianfang .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2021, 32 (03) :1067-1081
[7]   Deep Feature Extraction and Classification of Hyperspectral Images Based on Convolutional Neural Networks [J].
Chen, Yushi ;
Jiang, Hanlu ;
Li, Chunyang ;
Jia, Xiuping ;
Ghamisi, Pedram .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2016, 54 (10) :6232-6251
[8]   An Efficient Implementation of Convolutional Neural Network With CLIP-Q Quantization on FPGA [J].
Cheng, Wei ;
Lin, Ing-Chao ;
Shih, Yun-Yang .
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS, 2022, 69 (10) :4093-4102
[9]   Mixed-precision quantized neural networks with progressively decreasing bitwidth [J].
Chu, Tianshu ;
Luo, Qin ;
Yang, Jie ;
Huang, Xiaolin .
PATTERN RECOGNITION, 2021, 111
[10]  
Deng Li., 2012, Signal Processing Magazine, IEEE, V29, P141, DOI DOI 10.1109/MSP.2012.2211477