Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking

Chenghuan Liu, Du Huynh, Mark Reynolds

Research output: Chapter in Book/Conference paperConference paper

Abstract

Vehicle re-identification (Re-ID) plays a key role in many smart traffic management systems. Re-identifying a vehicle can be very challenging because the differences in visual appearances between pairs of vehicles are sometimes extremely subtle if they have the same colour and the same model. Given an image of a vehicle, most existing techniques adopt a global feature representation where details may be ignored. In this paper, we propose an Self-Attention Stair Feature Fusion model to learn the discriminative features for vehicle Re-ID. The model is designed to extract multi-level features in order to capture as much small details as possible. We also propose a Temporal Bayesian Re-Ranking method to exploit the spatial-temporal information in the vehicles’ travel patterns. Our algorithm has been tested against state-of-the-art techniques on popular benchmarks. The results show that our algorithm outperforms other state-of-the-art techniques by a large margin.
Original languageEnglish
Title of host publication International Joint Conference on Neural Networks (IJCNN)
Place of PublicationBudapest
PublisherIEEE, Institute of Electrical and Electronics Engineers
Number of pages8
DOIs
Publication statusPublished - 2019
EventInternational Joint Conference on Neural Networks - InterContinental Budapest Hotel, Budapest, Hungary
Duration: 14 Jul 201919 Jul 2019
https://www.ijcnn.org/

Conference

ConferenceInternational Joint Conference on Neural Networks
Abbreviated titleIJCNN 2019
CountryHungary
CityBudapest
Period14/07/1919/07/19
OtherThe 2019 International Joint Conference on Neural Networks (IJCNN) will be held at the InterContinental Budapest Hotel in Budapest, Hungary on July 14-19, 2019. The conference is organized by the International Neural Network Society (INNS) in cooperation with the IEEE Computational Intelligence Society, and is the premier international meeting for researchers and other professionals in neural networks and related areas. It will feature invited plenary talks by world-renowned speakers in the areas of neural network theory and applications, computational neuroscience, robotics, and distrbuted intelligence. In addition to regular technical sessions with oral and poster presentations, the conference program will include special sessions, competitions, tutorials and workshops on topics of current interest.
Internet address

Fingerprint

Stairs
Fusion reactions
Color

Cite this

Liu, C., Huynh, D., & Reynolds, M. (2019). Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking. In International Joint Conference on Neural Networks (IJCNN) Budapest: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/IJCNN.2019.8852350
Liu, Chenghuan ; Huynh, Du ; Reynolds, Mark. / Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking. International Joint Conference on Neural Networks (IJCNN). Budapest : IEEE, Institute of Electrical and Electronics Engineers, 2019.
@inproceedings{cfb5535e680a4be79b4920a5f12f5b09,
title = "Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking",
abstract = "Vehicle re-identification (Re-ID) plays a key role in many smart traffic management systems. Re-identifying a vehicle can be very challenging because the differences in visual appearances between pairs of vehicles are sometimes extremely subtle if they have the same colour and the same model. Given an image of a vehicle, most existing techniques adopt a global feature representation where details may be ignored. In this paper, we propose an Self-Attention Stair Feature Fusion model to learn the discriminative features for vehicle Re-ID. The model is designed to extract multi-level features in order to capture as much small details as possible. We also propose a Temporal Bayesian Re-Ranking method to exploit the spatial-temporal information in the vehicles’ travel patterns. Our algorithm has been tested against state-of-the-art techniques on popular benchmarks. The results show that our algorithm outperforms other state-of-the-art techniques by a large margin.",
author = "Chenghuan Liu and Du Huynh and Mark Reynolds",
year = "2019",
doi = "10.1109/IJCNN.2019.8852350",
language = "English",
booktitle = "International Joint Conference on Neural Networks (IJCNN)",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Liu, C, Huynh, D & Reynolds, M 2019, Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking. in International Joint Conference on Neural Networks (IJCNN). IEEE, Institute of Electrical and Electronics Engineers, Budapest, International Joint Conference on Neural Networks, Budapest, Hungary, 14/07/19. https://doi.org/10.1109/IJCNN.2019.8852350

Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking. / Liu, Chenghuan; Huynh, Du; Reynolds, Mark.

International Joint Conference on Neural Networks (IJCNN). Budapest : IEEE, Institute of Electrical and Electronics Engineers, 2019.

Research output: Chapter in Book/Conference paperConference paper

TY - GEN

T1 - Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking

AU - Liu, Chenghuan

AU - Huynh, Du

AU - Reynolds, Mark

PY - 2019

Y1 - 2019

N2 - Vehicle re-identification (Re-ID) plays a key role in many smart traffic management systems. Re-identifying a vehicle can be very challenging because the differences in visual appearances between pairs of vehicles are sometimes extremely subtle if they have the same colour and the same model. Given an image of a vehicle, most existing techniques adopt a global feature representation where details may be ignored. In this paper, we propose an Self-Attention Stair Feature Fusion model to learn the discriminative features for vehicle Re-ID. The model is designed to extract multi-level features in order to capture as much small details as possible. We also propose a Temporal Bayesian Re-Ranking method to exploit the spatial-temporal information in the vehicles’ travel patterns. Our algorithm has been tested against state-of-the-art techniques on popular benchmarks. The results show that our algorithm outperforms other state-of-the-art techniques by a large margin.

AB - Vehicle re-identification (Re-ID) plays a key role in many smart traffic management systems. Re-identifying a vehicle can be very challenging because the differences in visual appearances between pairs of vehicles are sometimes extremely subtle if they have the same colour and the same model. Given an image of a vehicle, most existing techniques adopt a global feature representation where details may be ignored. In this paper, we propose an Self-Attention Stair Feature Fusion model to learn the discriminative features for vehicle Re-ID. The model is designed to extract multi-level features in order to capture as much small details as possible. We also propose a Temporal Bayesian Re-Ranking method to exploit the spatial-temporal information in the vehicles’ travel patterns. Our algorithm has been tested against state-of-the-art techniques on popular benchmarks. The results show that our algorithm outperforms other state-of-the-art techniques by a large margin.

U2 - 10.1109/IJCNN.2019.8852350

DO - 10.1109/IJCNN.2019.8852350

M3 - Conference paper

BT - International Joint Conference on Neural Networks (IJCNN)

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Budapest

ER -

Liu C, Huynh D, Reynolds M. Urban Area Vehicle Re-Identification With Self-Attention Stair Feature Fusion and Temporal Bayesian Re-Ranking. In International Joint Conference on Neural Networks (IJCNN). Budapest: IEEE, Institute of Electrical and Electronics Engineers. 2019 https://doi.org/10.1109/IJCNN.2019.8852350