A Privacy-Aware Mobile Object Detection Tool for Urban Surveys
For urban planning, the tracking of pedestrian flows is a valuable tool. However, existing solutions either rely on labor-intensive manual counting, are prohibitively expensive for short-term analysis, or raise significant privacy concerns by sending image data to a central service. Consequently, this class of data has mostly been out of reach for interested citizens.
My master’s thesis is an app designed to turn a wide range of low-cost Android phones into a privacy-respecting pedestrian tracking tool. All image analysis is done locally on the device.
The whole document: Masters Thesis