Publication Details
Density-Based Vehicle Counting with Unsupervised Scale Selection
Špaňhel Jakub, Ing., Ph.D. (DCGM)
Bartl Vojtěch, Ing., Ph.D. (DCGM)
Juránek Roman, Ing., Ph.D. (DCGM)
Herout Adam, prof. Ing., Ph.D. (DCGM)
Deep learning; Density estimation; Object counting; Traffic surveillance; Vehicle
counting
A significant hurdle within any counting task is the variance in scale of the
objects to be counted. While size changes of some extent can be induced by
perspective distortion, more severe scale differences can easily occur, e.g. in
case of images taken by a drone from different elevations above the ground. The
aim of our work is to overcome this issue by leveraging only lightweight dot
annotations and a minimum level of training supervision. We propose
a modification to the Stacked Hourglass network which enables the model to
process multiple input scales and to automatically select the most suitable
candidate using a quality score. We alter the training procedure to enable
learning of the quality scores while avoiding their direct supervision, and thus
without requiring any additional annotation effort. We evaluate our method on
three standard datasets: PUCPR+, TRANCOS and CARPK. The obtained results are on
par with current state-of-the-art methods while being more robust towards
significant variations in input scale.
@inproceedings{BUT168491,
author="Petr {Dobeš} and Jakub {Špaňhel} and Vojtěch {Bartl} and Roman {Juránek} and Adam {Herout}",
title="Density-Based Vehicle Counting with Unsupervised Scale Selection",
booktitle="Digital Image Computing: Techniques and Applications 2020",
year="2020",
pages="1--8",
publisher="Institute of Electrical and Electronics Engineers",
address="Melbourne",
doi="10.1109/DICTA51227.2020.9363401",
isbn="978-1-7281-9108-9",
url="http://www.dicta2020.org/wp-content/uploads/2020/09/22_CameraReady.pdf"
}