Vehicle Logo Detection and Classification using Discriminative Pixel-patches Sparse Coding

DOI

10.2991/icmia-16.2016.58How to use a DOI?

Keywords

Linear regression, Sparse coding,Classification,SVM.

Abstract

Sparse coding has been successfully applied to many numerical object detection and classification problems in recent years. However, sparse coding has rather slow convergence rates on some optimization problems. In this article, an online framework for fast and classification of different vehicles logo in a natural cluttered background is proposed. Our basic idea is that the better a detection is for vehicle logo by two stages logo detection. Two techniques are applied to achieve it: making Pixel-patches Sparse Coding (PSC) to help suppress noise and enhance translation invariance of sparse representation. PSC describes the local appearance and encodes the each component relative position. The process starts with license plates localization using linear 3-channel pixels regression, followed by vehicle logo classification. This is achieved first computing the structural sparse coding and then looking for highly structured areas which usually correspond to the license plate and log areas. Detected areas are then fed to the classifier operating in the multi-class online structural SVM. The proposed detection and classification algorithm runs in real-time and performs favorably against state-of-the-art algorithms on challenging sequences in terms of efficiency, accuracy and robustness.

Copyright

© 2016, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Ouyang Yi
PY  - 2016/11
DA  - 2016/11
TI  - Vehicle Logo Detection and Classification using Discriminative Pixel-patches Sparse Coding
BT  - Proceedings of the 2016 5th International Conference on Measurement, Instrumentation and Automation (ICMIA 2016)
PB  - Atlantis Press
SP  - 329
EP  - 333
SN  - 1951-6851
UR  - https://doi.org/10.2991/icmia-16.2016.58
DO  - 10.2991/icmia-16.2016.58
ID  - Yi2016/11
ER  -