Patterns and predictability of visual attention in different street types : An eye tracking study exploring the predictability of the distribution of  human visual  attention based on the spatial arrangements of buildings in a two-dimensional plan.

Streets are the most resilient elements in a city. While buildings are replaced and property borders redrawn, the streets remain. The primary means of planning new streets and the built environment surrounding it is by the creation of a two-dimensional plan - in a Swedish context called a detailed development plan. The two-dimensional plan is sometimes criticized for its inability to take the three-dimensional world into account and thus its inability to predict the actual outcome of the plan. To address this critique and provide additional understanding for this planning device, this paper empirically explores if the distribution of visual attention can be predicted from a two-dimensional plan. Visual attention is explored due to sight being the most prominent means used by humans to gather information in the urban environment. Fifteen subjects performed a simple, computerized eye-tracking experiment, whereby their visual attention was measured as they looked at 40 images on a monitor. The 40 images depicts four different street types: blocked, open, curved and angled, each of which was represented with a simple figure image, as per interpretation from a two-dimensional plan, as well as a photograph of the same environment, representing the actual outcome of the plan. The results show that the simple figure images have effective predictive capabilities, as the distribution of visual attention exhibited a similar pattern in both the figure image and the photograph. The results also show that different patterns of visual attention are evoked by each of the four different street types. In sum, the results indicate that two-dimensional plans are able to predict the future three-dimensional outcome of a given plan in terms of visual attention. These indications are valuable for planners, architects, engineers and decision-makers when planning for new urban environments. The results are also valuable for understanding human perception of streets in a wider context