Abstract

Multiple object tracking (MOT) in unmanned aerial vehicle (UAV) videos has attracted attention. Because of the observation perspectives of UAV, the object scale changes dramatically and is relatively small. Besides, most MOT algorithms in UAV videos cannot achieve real-time due to the tracking-by-detection paradigm. We propose a feature-aligned attention network (FAANet). It mainly consists of a channel and spatial attention module and a feature-aligned aggregation module. We also improve the real-time performance using the joint-detection-embedding paradigm and structural re-parameterization technique. We validate the effectiveness with extensive experiments on UAV detection and tracking benchmark, achieving new state-of-the-art 44.0 MOTA, 64.6 IDF1 with 38.24 frames per second running speed on a single 1080Ti graphics processing unit.

© 2022 Chinese Laser Press

Mutation detection dual correlation filter with an object-awareness module for real-time target tracking

Baiheng Cao, Xuedong Wu, Yaonan Wang, Zhiyu Zhu, Huilin Ge, Runbang Liu, and Mengquan Liang
J. Opt. Soc. Am. A 39(9) 1712-1722 (2022)

MAFFNet: real-time multi-level attention feature fusion network with RGB-D semantic segmentation for autonomous driving

Tongfei Lv, Yu Zhang, Lin Luo, and Xiaorong Gao
Appl. Opt. 61(9) 2219-2229 (2022)

Real-time optical holographic tracking of multiple objects

Tien-Hsin Chao and Hua-Kuang Liu
Appl. Opt. 28(2) 226-231 (1989)

Camera-guided real-time laser ranging for multi-UAV distance measurement

Denis Ojdanić, Benjamin Gräf, Andreas Sinn, Han Woong Yoo, and Georg Schitter
Appl. Opt. 61(31) 9233-9240 (2022)

Object tracking mask-based NLUT on GPUs for real-time generation of holographic videos of three-dimensional scenes

M.-W. Kwon, S.-C. Kim, S.-E. Yoon, Y.-S. Ho, and E.-S. Kim
Opt. Express 23(3) 2101-2120 (2015)