Understanding the role of image recognition in mobile tour guides

Users of mobile tour guides often express a strong desire for the system to be able to provide information on arbitrary objects they encounter during their visit - akin to pointing to a building or attraction and saying "what's that ?" to a human tour guide. This paper reports on a field study in which we investigated user reaction to the use of digital image capture and recognition to support such functionality. Our results provide an insight into usage patterns and likely user reaction to mobile tour guides that use digital photography for real-time object recognition. These results include the counter-intuitive observation that a significant class of users appear happy to use image recognition even when this is a more complex, lengthy and error-prone process than traditional solutions. Careful analysis of user behavior during the field trails also provides evidence that it may be possible to classify tourists according to the methods by which they prefer to acquire information about tourist attractions in their vicinity. If shown to be generally true these results have important implications for designers of future mobile tour guide systems.