TY - GEN
T1 - Hybrid Optimization of Target Detection on Embedded Platforms for Real Time Applications
AU - Zhang, Xinchen
AU - Sun, Wangchao
AU - Zhao, Yaodong
AU - Liao, Kaisheng
AU - Liu, Yilin
AU - Xu, Hongda
AU - Xiao, Zhuoling
AU - Yan, Bo
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Target detection has been widely used in fields such as intelligent security and autonomous driving. However, existing computationally heavy target detection algorithms based on deep learning can only work on GPU and CPU platforms, restricting the applications on edge devices with limited computational power. To address this issue, this paper proposes layer fusion and 16-bit fixed-point quantization on the YOLOv2-Tiny algorithm to reduce the computational complexity of target detection algorithms. Furthermore, the data transmission efficiency is optimized by using ping-pong butter and multi-channel methods. To reduce FPGA resource consumption, the neural network is split into convolution, accumulation, pooling, and address mapping modules. The proposed system has been successfully implemented on the Xilinx Zynq-XC7Z035 platform, using only 47% of BRAM resources and 18% of DSP resources.
AB - Target detection has been widely used in fields such as intelligent security and autonomous driving. However, existing computationally heavy target detection algorithms based on deep learning can only work on GPU and CPU platforms, restricting the applications on edge devices with limited computational power. To address this issue, this paper proposes layer fusion and 16-bit fixed-point quantization on the YOLOv2-Tiny algorithm to reduce the computational complexity of target detection algorithms. Furthermore, the data transmission efficiency is optimized by using ping-pong butter and multi-channel methods. To reduce FPGA resource consumption, the neural network is split into convolution, accumulation, pooling, and address mapping modules. The proposed system has been successfully implemented on the Xilinx Zynq-XC7Z035 platform, using only 47% of BRAM resources and 18% of DSP resources.
KW - target detection
KW - YOLOv2-Tiny
KW - Zynq
UR - https://www.scopus.com/pages/publications/85136329765
U2 - 10.1109/ICET55676.2022.9825328
DO - 10.1109/ICET55676.2022.9825328
M3 - Conference contribution
AN - SCOPUS:85136329765
T3 - 2022 IEEE 5th International Conference on Electronics Technology, ICET 2022
SP - 1136
EP - 1141
BT - 2022 IEEE 5th International Conference on Electronics Technology, ICET 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 5th IEEE International Conference on Electronics Technology, ICET 2022
Y2 - 13 May 2022 through 16 May 2022
ER -