Agentic data science with the Data Agent Kit
cloud.google.com·6h ago
Data Science
Simon Willison's Blog·13d ago·5 min read
Photo Scrubber — local face blur & metadata removal
TL;DR
An experimental tool called Photo Scrubber was developed to automatically identify and blur faces in photographs, addressing privacy concerns when sharing images of strangers. It utilizes Google's MediaPipe library and the BlazeFace model for face detection.
✦ Why It Matters
Use Photo Scrubber to protect individuals' privacy when sharing images online.
Key Takeaways
- Implement face detection in your applications using MediaPipe.
- Explore WebAssembly for performance improvements in web tools.
- Consider privacy implications when sharing images of people.
- Utilize automated tools to enhance user privacy in digital content.
Full Summary
Photo Scrubber was created in response to privacy concerns about sharing images of individuals without their consent. The tool employs Google's MediaPipe C++ library, which is compiled to WebAssembly, along with the BlazeFace face detection model to identify faces in photographs.
Once detected, the tool automatically blurs these faces, ensuring that identifiable features are obscured. This approach allows users to share images of events, like protests, without compromising the privacy of individuals captured in the frame.
The integration of advanced face detection technology makes the tool efficient and user-friendly. As a result, it provides a practical solution for those looking to maintain privacy while documenting public events.