Higher ordered statistics-based bipolarity measure for detecting pedestrian crossing from visual signals

Abstract

The design and development of low cost portable device to enhance the mobility of visually impaired or blind people remains elusive and difficult to achieve. One of the key research problems in this context is the robust detection of pedestrian crossing to improve the safety and mobility of blind people. This paper details the development of detecting pedestrian crossing from images using the bipolar nature (alternate black and white strips) of the pedestrian crossing. A bipolarity measure based on higher order statistics (HOS) is introduced for the detection of pedestrian crossing. In particular, skewness is incorporated to define HOS-based bipolarity measure. Empirical analyses suggest that the HOS-based bipolarity measure performed better compared to the pedestrian crossing detection using the second ordered statistics.

Publication Title

Proceedings of the 2008 International Conference on Image Processing, Computer Vision, and Pattern Recognition, IPCV 2008

This document is currently not available here.

Share

COinS