Towards building an automated system for describing indoor floor maps for individuals with visual impairment

The ability to navigate is an integral part of everyday life, which is one of the major challenges faced by individuals with visual impairment. While navigating indoors is relatively safer, it is a stressful exercise if the building is unfamiliar, especially since typical means like canes, seeing-dogs and GPS devices do not really help. This article presents an automated approach that aids a visually impaired individual in obtaining information from a floor map, prior to visiting large buildings like a library. The key is to automatically produce verbal descriptions from a floor map and use that to help the person to form a cognitive map of the building to be visited. The proposed method is fully automated, which only requires the user to input the name of the building/establishment. The system then searches the Web for a floor map. If a map is found, the map is downloaded and used by the system to generate a verbal description giving an overview of the layout of the building, including entrances/exits and other important landmarks inside the building. A prototype system was built and evaluated with a group of 18 visually impaired people. Feedback obtained from the visually impaired participants in our experiments suggests that the proposed approach is an effective aid in helping the users forming an approximate cognitive map of the location before visiting the location.