Portrait of a City is a series of works representing a geographical identity of a city by the face generated by averaging Google satellite maps detected by a face detection algorithm.
Elements constituting our image of the city are complex. American urban planner Kevin Lynch suggests in his book, The Image of the City, there are five elements that shape our image of the city: paths, edges, districts, nodes, and landmarks. This project attempts to simply visualise the image of the city by translating those elements into one face.
This project also attempts to examine the difference between human vision and computer vision. Computer vision is modeled after human biology, but it sees human faces differently from the way humans do. The algorithm detects lots of landscapes as human faces, which are difficult for humans to recognise as human faces.
At the same time, what the algorithm judges as human faces is limited by faces included in the training data set for machine learning. Faces generated from landscapes paradoxically reveal what the algorithm has learned as features of the human face.