Certains contenus de cette application ne sont pas disponibles pour le moment.
Si cette situation persiste, veuillez nous contacter àObservations et contact
1. (WO2019005825) PROCÉDÉ ET SYSTÈME POUR METTRE EN ŒUVRE UNE AUTHENTIFICATION D'IMAGE POUR AUTHENTIFIER DES PERSONNES OU DES ARTICLES
Note: Texte fondé sur des processus automatiques de reconnaissance optique de caractères. Seule la version PDF a une valeur juridique

WHAT IS CLAIMED IS:

1. A method, comprising:

receiving, with a computing system, an image from an optical sensor device; identifying, with the computing system, one or more image regions from the received image;

analyzing, with the computing system, each of the one or more image regions to identify one or more spatial relationships amongst at least one of pixels or groups of pixels in each image region;

comparing, with the computing system, each identified spatial relationship amongst the at least one of pixels or groups of pixels in each image region with a plurality of spatial relationships amongst the at least one of pixels or groups of pixels that are characteristic of particular image artifacts; and generating, with the computing system, at least one of one or more

authenticity values or one or more results for the image based at least in part on results of the analysis and the comparison.

2. The method of claim 1, wherein the optical sensor device comprises at least one of an image capture device, a video capture device, a digital camera, a cell phone camera, a laptop computer camera, a tablet computer camera, a webcam, a security camera, a closed-circuit camera, a doorbell camera, an intercom camera, a drone-mounted camera, or a vehicle-mounted camera.

3. The method of claim 1, wherein the image received from the optical sensor device comprises an image extracted from a video stream that is received from the optical sensor device.

4. The method of claim 1, further comprising:

storing the received image in a data storage device, wherein storing the

received image comprises storing, in the data storage device, raw image data as captured by the optical sensor device, without aligning, cropping, filtering, resizing, compressing, or performing other image processing on the raw image data.

5. The method of claim 1, wherein identifying the one or more image regions from the received image comprises extracting, with the computing system, one or more image regions from the received image.

6. The method of claim 1, wherein identifying the one or more image regions comprises selecting, with the computing system, one of an arbitrary or a fixed number of image regions, wherein selecting the one of the arbitrary or the fixed number of image regions comprises one of using one or more machine learning techniques, using random selection techniques, or using image heuristics.

7. The method of claim 1, wherein identifying the one or more spatial relationships amongst the at least one of pixels or groups of pixels in each image region comprises:

obtaining, with the computing system, one or more authentic images and one or more inauthentic images, wherein the one or more authentic images each comprises image data of one of an actual person or an actual item as captured directly by the optical sensor device, and wherein the one or more inauthentic images each comprises image data of previously captured images of one of a person or an item;

extracting, with the computing system, one or more image regions from each of the one or more authentic images and from each of the one or more inauthentic images;

identifying, with the computing system, one or more spatial relationships of the at least one of pixels or groups of pixels that distinguish between authentic images and inauthentic images, using machine learning techniques; and

storing, with the computing system, the identified one or more spatial

relationships of the at least one of pixels or groups of pixels in a data storage device.

8. The method of claim 7, wherein the obtained one or more authentic images and the obtained one or more inauthentic images are captured by different types or models of optical sensor devices.

9. The method of claim 7, wherein analyzing each of the one or more image regions comprises applying, with the computing system, a layer of

convolutional kernels against each image region, wherein each convolutional kernel or combination of kernels corresponds to at least one of micro textures or other patterns that are indicative of differences between authentic images and inauthentic images.

10. The method of claim 9, wherein the micro textures comprise at least one of natural texture of skin on a human face from an image that is as captured directly by the optical sensor device, natural texture of a material of an item from an image that is as captured directly by the optical sensor device, a moire pattern from image capture of an image of a person, a moire pattern from image capture of an image of an item, an image artifact from a printed photograph of a person, an image artifact from a printed photograph of an item, a compression artifact from display of a photograph of a person, a compression artifact from display of a photograph of an item, a compression artifact from a screen display of an image of a person, a compression artifact from a screen display of an image of an item, or an image artifact resulting from image capture by one of particular types of optical sensor devices.

11. The method of claim 9, wherein analyzing each of the one or more image regions further comprises processing, with the computing system and using a max pooling filter, a response from each convolutional kernel or combination of kernels to determine whether a given convolutional filter has detected a signal corresponding to a micro texture of one or more convolutional kernels.

12. The method of claim 11, wherein analyzing each of the one or more image regions further comprises passing outputs of the max pooling filter into an inner product layer, and identifying, with the inner product layer, a linear combination of the response from each convolutional kernel or combination of kernels and generating, with the inner product layer, a linearly optimal score value corresponding to a likelihood that the image constitutes an authentic image.

13. The method of claim 1, further comprising:

assigning, with the computing system, a weighted likelihood value to each of the one or more image regions based on strength of the identified one or

more spatial relationships amongst the at least one of pixels or groups of pixels in each image regions;

wherein generating the at least one of the one or more authenticity values or the one or more results are further based at least in part on the assigned weighted likelihood value.

14. The method of claim 1, wherein the at least one of the one or more authenticity values or the one or more results for the image are indicative of a likelihood of authenticity of the image, and wherein the at least one of the one or more authenticity values or the one or more results for the image are further based at least in part on one or more aggregation techniques of weighted likelihood values across all analyzed image regions.

15. The method of claim 1, wherein the at least one of the one or more authenticity values or the one or more results for the image are indicative of a likelihood of authenticity of the image, and wherein the at least one of the one or more authenticity values or the one or more results for the image are further based at least in part on one or more aggregation techniques of weighted likelihood values across a subset of analyzed image regions.

16. An apparatus, comprising:

at least one processor; and

a non-transitory computer readable medium communicatively coupled to the at least one processor, the non-transitory computer readable medium having stored thereon computer software comprising a set of instructions that, when executed by the at least one processor, causes the apparatus to: receive an image from an optical sensor device;

extract one or more image regions from the received image;

analyze each of the one or more image regions to identify one or more spatial relationships amongst the at least one of pixels or groups of pixels in each image region;

compare each identified spatial relationship amongst the at least one of pixels or groups of pixels in each image region with a plurality of spatial relationships amongst the at least one of pixels or groups of pixels that are characteristic of particular image artifacts; and

generate at least one of one or more authenticity values or one or more results for the image based at least in part on results of the analysis and the comparison.

17. The apparatus of claim 16, wherein the optical sensor device comprises at least one of an image capture device, a video capture device, a digital camera, a cell phone camera, a laptop computer camera, a tablet computer camera, a webcam, a security camera, a closed-circuit camera, a doorbell camera, an intercom camera, a drone-mounted camera, or a vehicle-mounted camera.

18. A system, comprising:

a computing system, comprising:

at least one first processor; and

a first non-transitory computer readable medium communicatively coupled to the at least one first processor, the first non-transitory computer readable medium having stored thereon computer software comprising a first set of instructions that, when executed by the at least one first processor, causes the computing system to: receive an image from an optical sensor device;

extract one or more image regions from the received image; analyze each of the one or more image regions to identify one or more spatial relationships amongst the at least one of pixels or groups of pixels in each image region;

compare each identified spatial relationship amongst the at least one of pixels or groups of pixels in each image region with a plurality of spatial relationships amongst the at least one of pixels or groups of pixels that are characteristic of particular image artifacts;

generate at least one of one or more authenticity values or one or more results for the image based at least in part on results of the analysis and the comparison; and

sending the generated at least one of the one or more

authenticity values or the one or more results to a display device;

the display device, comprising:

a display screen;

at least one second processor; and

a second non-transitory computer readable medium communicatively coupled to the at least one second processor, the second non- transitory computer readable medium having stored thereon computer software comprising a second set of instructions that, when executed by the at least one second processor, causes the display device to:

receive the generated at least one of the one or more

authenticity values or the one or more results; and display the generated at least one of the one or more

authenticity values or the one or more results on the display screen.

19. The system of claim 18, wherein the computing system comprises one of a tablet computer, a smart phone, a mobile phone, a laptop computer, a desktop computer, a gaming console, a media player, a server computer over a network, or a cloud-based computing system over a network.