Google open-sources two internal privacy-enhancing technologies – Engadget 中文版

as a driving forceprotection operation” plan, Google has decided to combine two internalPrivacy Enhancement Technology(PET) is officially open source, hoping to “technically ensure privacy and security” through changes in “data processing methods, time, and location”.Of the two technologies, the Magritte (now available in Github Published above) uses machine learning to detect objects and apply blur effects as soon as they appear in the frame. “Magritte is especially useful for videojournalists who want to provide additional guarantees of privacy,” Google wrote in a blog post. “By using this open source code, videographers can rely on the underlying machine learning precision detection, saving time on blurry objects.”

And another “Fully Homomorphic Encryption Transpiler” (Fully Homomorphic Encryption Transpiler) allows developers to use encrypted data to perform calculations without involving personally identifiable information. Google believes that it can be used in financial, medical and government fields. help. “In these industries, strong security safeguards around the handling of sensitive data are critical,” Google said. According to them, privacy-enhancing technologies are currently undergoing a transition from academic to mainstream. Both the U.S. and U.K. governments have pushed hard this year around solutions to financial crime and public health emergencies, and the trend is expected to extend to more diverse categories in the future.