Automatic Signboard Detection and Semi-Automatic Ground Truth Generation

Abstract

Data-driven object detection techniques are widely applied to a variety of practical areas (i.e., automatic robots, self-driving vehicles, defect detection, and face detection). Nowadays, many research projects have been proposed to improve the accuracy of computer vision applications. In this paper, we propose an automatic signboard detection method and a semi-automatic ground truth generation method to help visually impaired people walk on streets in Taiwan. We consider that when visually impaired people walk down the street, they may be interested in certain stores. Therefore, we collect images of 12 kinds of the most popular stores in people's daily lives. The collected street images number over 5 million from 6 major cities in Taiwan; however, only about 1% of images contain a signboard. We propose a hierarchical object detection module to pre-label uncertain samples. Based on this module, semi-automatic ground truth generation can be achieved.