Systems and methods for processing and handling privacy-sensitive image data
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for handling data including images with privacy-sensitive data. In one aspect, a method may include recognizing privacy-sensitive sub-image(s) in an acquired or captured image. The sub-image(s) can be included in a second image that is encrypted. The acquired image can be changes by obscuring the privacy-sensitive sub-image(s) of the acquired image so as not to reveal any personal identifiable information.
1. A method comprising:
obtaining, at one or more computing devices, an original image;
identifying, by the one or more computing devices, one or more privacy-sensitive areas in the obtained image;
creating, by the one or more computing devices, a privacy-sensitive image from the obtained image, the privacy-sensitive image comprising the one or more identified privacy-sensitive areas of the obtained original image;
encrypting, by the one or more computing devices, the privacy-sensitive image;
modifying, by the one or more computing devices, the image by obfuscating the one or more identified privacy-sensitive areas of the obtained original image;
storing, by the one or more computing devices, the modified image and the encrypted privacy-sensitive copy together; and
transmitting the stored together modified image and the encrypted privacy-sensitive copy to a trusted platform.
2. The method of claim 1 , wherein the modified image and the encrypted privacy-sensitive copy are stored together in a container structure.
3. The method of claim 1 , further comprising, storing, by the one or more computing devices, privacy-sensitive location data with the modified image and the privacy-sensitive copy, the privacy-sensitive location data indicating locations of the one or more identified privacy-sensitive areas of the obtained image.
4. The method of claim 1 , wherein the privacy-sensitive areas of the obtained image comprise one or more facial images.
5. The method of claim 1 , wherein the privacy-sensitive areas of the obtained image comprise one or more license plate images.
6. The method of claim 1 , wherein identifying the one or more privacy-sensitive areas in the obtained image comprises applying a computer vision algorithm to the obtained image.
7. The method of claim 1 , wherein obfuscating the obtained image comprises applying a mask to the one or more identified privacy-sensitive areas of the obtained image.
8. The method of claim 7 , wherein applying a mask comprises pixelating or blurring the obtained image at the one or more identified privacy-sensitive areas of the obtained image.
9. A method for handling privacy-sensitive data and executed by one or more computing devices, the method comprising:
obtaining a modified image comprising one or more obfuscated privacy-sensitive areas;
obtaining an encrypted privacy-sensitive image associated with the obtained modified image comprising one or more unobfuscated privacy-sensitive images corresponding to the one or more privacy-sensitive areas of the associated obtained image;
decrypting the encrypted privacy-sensitive image;
generating a constructed image in a protected memory segment by using the privacy-sensitive image and the modified image, the constructed image being a version of the modified image with the one or more privacy-sensitive areas unobfuscated,
wherein the modified image and the encrypted privacy-sensitive image are part of a same file or container data structure so that the modified image and the encrypted privacy-sensitive image are obtained together.
10. The method of claim 9 , wherein the constructed image comprises a copy of a previous version of the modified image data.
11. The method of claim 9 , further comprising: allowing access to the constructed image in the protected memory segment by an application, wherein the application accesses the constructed image while operating in a protected memory mode; and removing access to the constructed image after the application has finished using with the constructed image.
12. The method of claim 11 , further comprising: deleting the constructed image from the protected memory; and deleting the decrypted privacy-sensitive image.
13. The method of claim 11 , wherein the application is a machine learning application, and the machine learning application accesses the constructed image through an API.
14. The method of claim 9 , wherein the protected memory segment is not accessible by a computing device.
15. One or more computing devices comprising one or more processors and a non-transitory computer-readable storage medium, including instructions that, when executed by the one or more processors, cause the one or more processors to:
obtain an image;
identify one or more privacy-sensitive areas in the obtained image; create a privacy-sensitive image from the obtained image, the privacy-sensitive image comprising the one or more identified privacy-sensitive areas of the obtained image;
encrypt the privacy-sensitive image; modify the image by obfuscating the one or more identified privacy-sensitive areas of the obtained image; and
store the modified image and the encrypted privacy-sensitive copy together; and
transmit the stored together modified image and the encrypted privacy-sensitive copy to a trusted platform.
16. One or more computing devices comprising one or more processors and a non-transitory computer-readable storage medium, including instructions that, when executed by the one or more processors, cause the one or more processors to:
obtain a modified image comprising one or more privacy-sensitive areas, the one or more privacy-sensitive areas being obfuscated;
obtain an encrypted privacy-sensitive image that is associated with the obtained modified image, the privacy-sensitive image comprising one or more unobfuscated image areas corresponding to the one or more privacy-sensitive areas of the associated obtained image;
decrypt the privacy-sensitive image;
generate a constructed image in a protected memory segment using the privacy-sensitive image and the modified image, the constructed image being a version of the modified image with the one or more privacy-sensitive areas unobfuscated,
wherein the modified image and the encrypted privacy-sensitive image are part of a same file or container data structure so that the one or more processors obtain the modified image and the encrypted privacy-sensitive image together.
17. The one or more computing devices of claim 16 , wherein the one or more processors are further caused to: allow access to the constructed image in the protected memory segment by a machine learning application; and removing access to the constructed image after the machine learning application has finished using with the constructed image.