ML-CB

In this paper, we develop a new privacy enhancing tool: ML-CB—a means of using distinguishable pictorial information combined with underlying website source code to produce accurate and robust machine learning classifiers able to discern fingerprinting (i.e., surreptitious tracking) from non-fingerprinting canvas-based actions.

The data introduced in the paper is collected by scraping roughly half a million websites using a custom Google Chrome extension storing information related to the canvas.

Source: ML-CB: Machine Learning Canvas Block