BScanner: A Crowdsourcing Platform for Constructing Accessibility Maps to Support Multiple Participation Types

Abstract

Accessibility maps enable impaired/elderly people to move more smoothly. However, very few cases satisfy both the accuracy and coverage required, due to the high cost of physically auditing roads. Though crowdsourcing approaches could solve this problem, existing studies rely on volunteers who are highly motivated and have enough free time. In this paper, we propose a crowdsourcing platform for constructing accessibility maps that supports multiple participation types: manual reporting for people who are highly motivated and have free time, walking for people who are highly motivated but do not have free time, and game playing for people who are less motivated but have free time. This design allows people to select a suitable way to participate, depending on their motivation and time. We have developed a prototype system by integrating deep learning techniques, a game design theory, and heatmap visualization.