Graphic data alteration to enhance online privacy
View Patent ↗A computer alters at least one recognizable metric or text in a digitally encoded photographic image by operating an alteration algorithm in response to user input data while preserving an overall aesthetic quality of the image and obscuring an identity of at least one individual or geographic location appearing in the image. An altered digitally-encoded photographic image prepared by the altering of the at least one recognizable metric or text in the image is stored in a computer memory. User feedback and/or automatic analysis may be performed to define parameter values of the alteration algorithm such that the alteration process achieves preservation of aesthetic qualities while obscuring an identity of interest.
1. A system configured to impair automated image analysis, comprising:
a processor coupled to a memory, the memory encoded with instructions that when executed by the processor, causes the system to:
analyze a digitally-encoded image comprising one or more words and at least one face;
alter selected facial recognition metrics in the at least one face using an alteration algorithm that identifies an alteration of the selected facial recognition metrics based on a value of
Δ
=
∑
1
to
M
∂
i
_
,
configured to substantially reduce risk of automatic recognition of the face while preserving an overall aesthetic quality of the image, wherein δi is a measure of metric alteration of each of a number M of the selected facial recognition metrics;
identify at least one word in the digitally-encoded image;
alter the digitally-encoded image so that the position of at least two characters in the word are reversed, and the first and last letter of the word are in a proper position; and
store in the memory, the altered digitally-encoded image.
2. The system of claim 1 , wherein the processor further causes the system to modify at least one parameter of the alteration algorithm in response to user feedback data.
3. The system of claim 1 , wherein
a median common feature set, different than the actual feature set, is used as a starting point in generating variations across a group of images.
4. The system of claim 1 , wherein the image altered is published in an electronic medium, user feedback is gathered with respect to the aesthetic quality of the image altered, and changes are made to the alteration algorithm based on the feedback.
5. The system of claim 1 , wherein the processor further causes the system to introduce random or quasi-random variability in altered pixels configured such that altered pixel values are not rigidly determined by adjoining unaltered pixels appearing in the image.
6. The system of claim 1 , wherein a median common facial feature set, different than an actual facial feature set of the at least one face, is used as a starting point in generating variations for the face across more than one image.
7. The system of claim 1 , further comprising modifying at least one parameter of the alteration algorithm in response to user input specifying a minimum confidence level.
8. A computer system configured to impair automated image analysis, comprising:
one or more hardware processors programmed, via executable code instructions, to implement:
analyzing a digital image;
identifying at least one object in the image;
altering at least one characteristic of the object in the image using an alteration algorithm that identifies an alteration of the selected object recognition metrics based on a value of
Δ
=
∑
1
to
M
∂
i
_
,
configured to substantially reduce risk of automatic recognition of the object when compared to other objects of the same kind while preserving an overall aesthetic quality of the image, wherein δi is a measure of metric alteration of each of a number M of the selected object recognition metrics.
9. The computer system of claim 8 , wherein the digital image comprises a face.
10. The computer system of claim 9 , wherein the at least one object is one of a vehicle or a building.
11. The computer system of claim 9 , wherein the at least one object is an indicator of a geographic location.
12. The computer system of claim 9 , wherein the at least one object is the face.
13. The computer system of claim 12 , wherein a median common facial feature set, different than an actual facial feature set of the at least one face, is used as a starting point in generating variations for the face across more than one image.
14. The computer system of claim 9 , where the object is a body.
15. The computer system of claim 9 , wherein the alteration algorithm creates random or quasi-random variability in altered pixel values such that the altered pixel values are not rigidly determined by adjoining unaltered pixels appearing in the image.
16. The computer system of claim 9 , where the image altered is published in an electronic medium, user feedback is gathered with respect to the aesthetic quality of the image altered, and changes are made to the alteration algorithm based on the feedback.
17. A computer system configured to impair automated face recognition analysis, comprising:
one or more hardware processors programmed, via executable code instructions, to implement:
analyzing at least two digital images, each containing the same person's face;
identifying the face in each image;
altering at least one characteristic of the face in at least one of the images so that automated face recognition techniques are rendered less effective in uniquely identifying the face; and
applying same alterations to the face in each of the other images.
18. The computer system of claim 17 , wherein the images are not analyzed at the same time.
19. The computer system of claim 17 , wherein a median common facial feature set, different than an actual facial feature set of the at least one face, is used as a starting point in generating variations for the face across a group of images.
20. The computer system of claim 17 , wherein the images altered using the same alterations are grouped based on text or tags in a contextual web page.