An Energy-Efficient Reconfigurable AI-Based Object Detection and Tracking Processor Supporting Online Object Learning

• Published in: IEEE solid-state circuits letters 2022, Vol.5, p.78-81
• Main Authors: Gong, Yuchuan, Zhang, Teng, Guo, Hongtao, Liu, Qingsong, Que, Luying, Jia, Conghan, Huang, Jiahui, Liu, Ye, Gan, Jiayan, Xie, Yuxiang, Zhou, Yong, Liu, Lili, Xiang, Xiaoqiang, Chang, Liang, Yan, Rui, Zhou, Jun
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Computer architecture; Control systems; Energy efficiency; Energy efficient; Engines; Learning; Meteorological satellites; Micromechanical devices; Microprocessors; neural network (NN); Neural networks; Object detection; object detection and tracking; Object recognition; processor; Random access memory; reconfigurable; Reconfiguration; Tracking
• ISSN: 2573-9603; 2573-9603
• DOI: 10.1109/LSSC.2022.3163478

This letter presents an energy-efficient reconfigurable AI-based object detection and tracking processor for smart drone/robot applications. Several techniques have been proposed to achieve high energy efficiency while supporting flexible object detection and tracking tasks with online object learning, including a reconfigurable object detection and tracking architecture with reconfigurable neural network (NN) engine, an online object learning architecture with shared NN inference and learning engine and automatic label generation engine, and a layer- and stride-aware NN computing technique. Compared with several state-of-the-art designs, the proposed design achieves better energy efficiency (2.13 mJ/frame), while supporting flexible object detection and tracking tasks with online object learning.