Zebra Crosswalk Region Detection and Localization Based on Deep Convolutional Neural Network

Abstract

It can be difficult for blinds and people with limited visual capabilities to find street intersections containing a crosswalk along with their accurate location. In this paper, a solution to this issue is proposed through a deep convolutional neural network (DCNN) architecture that automatically organizes several characteristics of zebra stripe crosswalks to support quick, accurate and reliable identification and detection of a crosswalk in an image. Proposed method uses Faster R-CNN Inception-v2 to identify and locate crosswalks, which has sparse convolutions on the same layer to reduce computational load while increasing accuracy. We focused on the single class – crosswalk, training the network with images of our own dataset combined with extracted image frames. To the best of our knowledge, proposed framework is the first to utilize deep architectures for crosswalk detection and localization from the street level view. It achieves an accuracy of 97.50% and is compared to previous method to show higher detection accuracy over recent works.