TY - GEN
T1 - Deep Learning-Based Vehicle Classification Using an Ensemble of Local Expert and Global Networks
AU - Lee, Jong Taek
AU - Chung, Yunsu
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/8/22
Y1 - 2017/8/22
N2 - Vehicle classification has been a challenging problem because of pose variations, weather / illumination changes, inter-class similarity and insufficient training dataset. With the help of innovative deep learning algorithms and large scale traffic surveillance dataset, we are able to achieve high performance on vehicle classification. In order to improve performance, we propose an ensemble of global networks and mixture of K local expert networks. It achieved a mean accuracy of 97.92%, a mean precision of 92.98%, a mean recall of 90.24% and a Cohen Kappa score of 96.75% on unseen test dataset from the MIO-TCD classification challenge.
AB - Vehicle classification has been a challenging problem because of pose variations, weather / illumination changes, inter-class similarity and insufficient training dataset. With the help of innovative deep learning algorithms and large scale traffic surveillance dataset, we are able to achieve high performance on vehicle classification. In order to improve performance, we propose an ensemble of global networks and mixture of K local expert networks. It achieved a mean accuracy of 97.92%, a mean precision of 92.98%, a mean recall of 90.24% and a Cohen Kappa score of 96.75% on unseen test dataset from the MIO-TCD classification challenge.
UR - http://www.scopus.com/inward/record.url?scp=85030254579&partnerID=8YFLogxK
U2 - 10.1109/CVPRW.2017.127
DO - 10.1109/CVPRW.2017.127
M3 - Conference contribution
AN - SCOPUS:85030254579
T3 - IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops
SP - 920
EP - 925
BT - Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017
PB - IEEE Computer Society
T2 - 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2017
Y2 - 21 July 2017 through 26 July 2017
ER -