BCN Perceptual Computing Lab Repository

Fast Traffic Sign Detection on greyscale images

Baró, Xavier and Vitrià, Jordi (2004) Fast Traffic Sign Detection on greyscale images. Recent Advances in Artificial Intelligence Research and Development. Frontiers in Artificial Intelligence and Applications, 113 . pp. 209-216. ISSN 978-1-58603-466-5

PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader


This paper describes a traffic sign detection framework for greyscale images. The system is a heterogeneous cascade classifier formed by a rectangle features cascade followed by specific filters for each shape. We define two types of filters: The first one is based on local changes of the gradient direction and the second one is based on the idea of radial symmetry and gives us the centre of circular shapes. The filters use the position and the aspect ratio of the detected features to differentiate between real objects and false alarms of the rectangle cascade. At the end we use the PCA to eliminate the remaining false alarms.

Item Type:Article
Subjects:Q Science > QA Mathematics > QA75 Electronic computers. Computer science
ID Code:21
Deposited By:Xavier Baró
Deposited On:03 Aug 2009 20:27
Last Modified:29 Nov 2010 17:52

Repository Staff Only: item control page