Scalable and parameterized dynamic time warping architecture for efficient vehicle re-identification
Guanbing Deng, Hanqing Zhou, Guangyu Yu, Zeyu Yan, Yu Hen Hu, Xiaowei Xu · 2017
Vehicle Re-identification is an effective way for calculation of travel time and origin-destination matrices, which is critical for traffic modeling and optimization. Recently, dynamic time warping (DTW) with magnetic signature has been a popular method for vehicle re-identification. However, with the increasing number of vehicles and the real-time requirement for travel time estimation, the calculation of DTW needs further acceleration. In this paper, we propose a scalable and parameterized architecture on reconfigurable fabrics for efficient vehicle re-identification. Parameterization is adopted to support various lengths of magnetic sequences. The experimental results with magnetic signature indicate that compared to the multi-core CPU-based implementation, our approach demonstrates over one order of magnitude on speedup and three orders of magnitude on energy-efficiency.