Publication Details
Vehicle Re-Identification and Multi-Camera Tracking in Challenging City-Scale Environment
Bartl Vojtěch, Ing., Ph.D. (DCGM FIT BUT)
Juránek Roman, Ing., Ph.D. (DCGM FIT BUT)
Herout Adam, prof. Ing., Ph.D. (DCGM FIT BUT)
vehicle re-identification, vehicle multi-camera tracking, city-scale environment, camera calibration, neural networks, nvidia ai city challenge
In our submission to the NVIDIA AI City Challenge, we address vehicle re-identification and vehicle multi-camera tracking. Our approach to vehicle re-identification is based on the extraction of visual features and aggregation of these features in the temporal domain to obtain a single feature descriptor for the whole observed track. For multi-camera tracking, we proposed a method for matching vehicles by the position of trajectory points in real-world space (linear coordinate system). Furthermore, we use CNN for the vehicle re-identification task to filter out false matches generated by proposed positional matching method for better results.
@INPROCEEDINGS{FITPUB12023, author = "Jakub \v{S}pa\v{n}hel and Vojt\v{e}ch Bartl and Roman Jur\'{a}nek and Adam Herout", title = "Vehicle Re-Identification and Multi-Camera Tracking in Challenging City-Scale Environment", pages = "150--158", booktitle = "2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)", series = "IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops", journal = "Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)", volume = 2019, number = 1, year = 2019, location = "Long Beach, US", publisher = "IEEE Computer Society", ISSN = "2160-7516", language = "english", url = "https://www.fit.vut.cz/research/publication/12023" }