Anonymization of facial images
A method facilitates the use of facial images through anonymization of facial images, thereby allowing people to submit their own facial images without divulging their identities. Original facial images are accessed and perturbed to generate synthesized facial images. Personal identities contained in the original facial images are no longer discernable from the synthesized facial images. At the same time, each synthesized facial image preserves at least some of the original attributes of the corresponding original facial image.
1. A computer-implemented method for generating synthesized facial images, the method comprising:
receiving an image;
determining the image is the original facial image by detecting a face of a live human subject in the image;
receiving a preselected attribute reflected in facial images;
determining the original facial image has the preselected attribute; and
perturbing the original facial image to generate a synthesized facial image, the synthesized facial image no longer recognizable as the live human subject but preserving at least part of the preselected attribute of the original facial image.
2. The computer-implemented method of claim 1 , wherein the synthesized facial image is no longer recognizable as the live human subject when a mutual information between a personal identity of the live human subject and the synthesized facial image falls below a threshold.
3. The computer-implemented method of claim 1 , wherein the synthesized facial image is no longer recognizable as the live human subject when a probability that humans can correctly identify the live human subject from the synthesized facial image is no greater than a threshold.
4. The computer-implemented method of claim 1 , wherein the facial image is captured under additional circumstances that probe for the preselected attribute.
5. The computer-implemented method of claim 1 , wherein the preselected attribute comprises at least a facial expression or emotional expression.
6. The computer-implemented method of claim 1 , further comprising outputting for presentation the synthesized facial image that is no longer recognizable as the live human subject.
7. A non-transitory computer readable medium containing instructions that, when executed by a processor, cause the processor to:
receive an image;
determine the image is the original facial image by detecting a face of a live human subject in the image;
receive a preselected attribute reflected in facial images;
determine the original facial image has the preselected attribute; and
perturb the original facial image to generate a synthesized facial image, the synthesized facial image no longer recognizable as the live human subject but preserving at least part of the preselected attribute of the original facial image.
8. The non-transitory computer readable medium of claim 7 , wherein the synthesized facial image is no longer recognizable as the live human subject when a mutual information between a personal identity of the live human subject and the synthesized facial image falls below a threshold.
9. The non-transitory computer readable medium of claim 7 , wherein the synthesized facial image is no longer recognizable as the live human subject when a probability that humans can correctly identify the live human subject from the synthesized facial image is no greater than a threshold.
10. The non-transitory computer readable medium of claim 7 , wherein the facial image is captured under additional circumstances that probe for the preselected attribute.
11. The non-transitory computer readable medium of claim 7 , wherein the preselected attribute comprises at least a facial expression or emotional expression.
12. The non-transitory computer readable medium of claim 7 , wherein the computer readable medium containing instructions that further cause the processor to output for presentation the synthesized facial image that is no longer recognizable as the live human subject.
13. A system for generating synthesized facial images, the system comprising:
a processor; and
a memory coupled to the processor and comprising instructions which, when executed by the processor, cause the system to:
receive an image;
determine the image is the original facial image by detecting a face of a live human subject in the image;
receive a preselected attribute reflected in facial images;
determine the original facial image has the preselected attribute; and
perturb the original facial image to generate a synthesized facial image, the synthesized facial image no longer recognizable as the live human subject but preserving at least part of the preselected attribute of the original facial image.
14. The system of claim 13 , wherein the synthesized facial image is no longer recognizable as the live human subject when a mutual information between a personal identity of the live human subject and the synthesized facial image falls below a threshold.
15. The system of claim 13 , wherein the synthesized facial image is no longer recognizable as the live human subject when a probability that humans can correctly identify the live human subject from the synthesized facial image is no greater than a threshold.
16. The system of claim 13 , wherein the facial image is captured under additional circumstances that probe for the preselected attribute.
17. The system of claim 13 , wherein the preselected attribute comprises at least a facial expression or emotional expression.