Pictures often contain faces of people even though this information is not needed at all for the purpose of the picture's processing. AI can help to identify faces and erase them from an image to allow for privacy-compliant downstream processing.
When hundreds or thousands of images are uploaded to a website or fed into a system it is almost impossible to control the pictures context on a manual level. Hand-picking pictures and masking people on it would be too expensive and tedious to erase them manually.
A face detection AI service an detect faces in images and deliver bounding boxes with pixel information for each face. This pixel information can then be used to modify the image by for example blurring out the respective areas or masking them.
Anonymized images can be used:
- for downstream processes that do not require personal identifiable information (IIR) to secure user privacy
- to increase user privacy
- to filter out images which contain people for regulated processes.