Qianhui Liang / Massachusetts Institute of Technology
In urban theory, urban form of an urban region well corresponds to its social and economic status. Certain type of urban form holds a high concentration of certain group of people. It is thus necessary to devel-op a better understanding of the relationship between physical features of the urban environment and social classes of residences. Figure-ground map is a simple, prevailing and precise representation of urban form in the field of urban study. Deep learning in computer vision en-ables such simple representation maps to be studied at a large scale. We propose to train a DCNN model to identify and uncover the inter-nal bridge between social class and urban form. Further, using hand-crafted visual features (building size, building density, etc.), which have significant impact in urban theory, we apply a random forest classifier to interpret how visual features are related with social class.