Method and system of using image recognition and geolocation signal analysis in the construction of a social media user identity graph
First Claim
Patent Images
1. A non-transitory computer readable medium having stored thereon computer-executable instructions, the computer executable instructions causing a processor to execute a method of creating an integrated, collated social media profile for a person which comprises:
- a) probing social media outlets and channels using one or more of social media application programming interfaces (APIs), data provided by third party data compilers, data aggregated directly from web crawlers, using selected person-specific search parameters and person specific context data to form a social media data stream;
b) from the social media data stream, identifying at least one photographic image from a first social media channel and at least one photographic image from a second social media channel;
c) comparing the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel via a plurality of image recognition steps, comprising for each image;
i) creating a 64 bit signature hash;
ii) converting the image down to 9×
8 pixels to remove unrequired details and to produce a low frequency 9×
8 image to ensure that different scaling and stretching will not affect a hash;
iii) converting the low frequency 9×
8 image into greyscale to simplify further computation;
iv) computing a greyscale colour difference between each pixel next to each other in each row of pixels, wherein each row will produce eight difference values for a total of 64 values;
v) creating a 64 bit hash wherein each bit is set to 0 or 1 depending on whether the respective left pixel is brighter than the right pixel and vi) comparing two images using Hamming distance to produce global image comparison output;
d) comparing the global image comparison output to a threshold to produce a set of resultant global images, wherein the set of resultant global images comprises at least one pair of photographic images from the first social media channel and the second social media channel of the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel having a global image comparison output less than the threshold;
e) comparing the set of resultant global images from the first social media channel and the set of resultant global images from the second social media channel via a plurality of image recognition steps, comprising for each image, i) identifying one or more features/keypoints;
ii) assigning descriptors to the one or more features/keypoints so that they can be recognized and compared to other images;
iii) weighing the one or more features/keypoints;
iv) matching the one or more features/keypoints;
f) from the social media data stream, i) identifying a plurality of geolocation data inputs related to the person and/or social media connections of the person;
ii) performing a centroid-based analysis on the plurality of geolocation data inputs to form clusters;
iii) classifying and filtering the clusters; and
iv) selecting one or more probable locations of the person based upon the clusters; and
g) determining and corroborating with confidence the photographic images and the geolocation data inputs which are assignable to the person and creating the integrated, collated social media profile, displayable on a dashboard, for the person, wherein the integrated, collated social media profile for the person comprises at least one matched photographic image of the set of resultant global images.
2 Assignments
0 Petitions
Accused Products
Abstract
A social media identity/construct is created for a person by locating social media profiles of the person at least by one or both of i) recognizing image(s) in one or more social media postings related to the person and ii) correlating geolocation data embedded in one or more social media postings related to the person or the person'"'"'s social media connections, analyzing, monitor social media outlets and channels to create a social media profile for the person.
51 Citations
20 Claims
-
1. A non-transitory computer readable medium having stored thereon computer-executable instructions, the computer executable instructions causing a processor to execute a method of creating an integrated, collated social media profile for a person which comprises:
-
a) probing social media outlets and channels using one or more of social media application programming interfaces (APIs), data provided by third party data compilers, data aggregated directly from web crawlers, using selected person-specific search parameters and person specific context data to form a social media data stream; b) from the social media data stream, identifying at least one photographic image from a first social media channel and at least one photographic image from a second social media channel; c) comparing the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel via a plurality of image recognition steps, comprising for each image;
i) creating a 64 bit signature hash;
ii) converting the image down to 9×
8 pixels to remove unrequired details and to produce a low frequency 9×
8 image to ensure that different scaling and stretching will not affect a hash;
iii) converting the low frequency 9×
8 image into greyscale to simplify further computation;
iv) computing a greyscale colour difference between each pixel next to each other in each row of pixels, wherein each row will produce eight difference values for a total of 64 values;
v) creating a 64 bit hash wherein each bit is set to 0 or 1 depending on whether the respective left pixel is brighter than the right pixel and vi) comparing two images using Hamming distance to produce global image comparison output;d) comparing the global image comparison output to a threshold to produce a set of resultant global images, wherein the set of resultant global images comprises at least one pair of photographic images from the first social media channel and the second social media channel of the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel having a global image comparison output less than the threshold; e) comparing the set of resultant global images from the first social media channel and the set of resultant global images from the second social media channel via a plurality of image recognition steps, comprising for each image, i) identifying one or more features/keypoints;
ii) assigning descriptors to the one or more features/keypoints so that they can be recognized and compared to other images;
iii) weighing the one or more features/keypoints;
iv) matching the one or more features/keypoints;f) from the social media data stream, i) identifying a plurality of geolocation data inputs related to the person and/or social media connections of the person;
ii) performing a centroid-based analysis on the plurality of geolocation data inputs to form clusters;
iii) classifying and filtering the clusters; and
iv) selecting one or more probable locations of the person based upon the clusters; andg) determining and corroborating with confidence the photographic images and the geolocation data inputs which are assignable to the person and creating the integrated, collated social media profile, displayable on a dashboard, for the person, wherein the integrated, collated social media profile for the person comprises at least one matched photographic image of the set of resultant global images. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A non-transitory computer readable medium having stored thereon computer-executable instructions, the computer executable instructions causing a processor to execute a method of creating an integrated, collated social media profile for a person which comprises:
-
a) probing social media outlets and channels using one or more of social media application programming interfaces (APIs), data provided by third party data compilers, data aggregated directly from web crawlers, using selected person-specific search parameters and person specific context data to form a social media data stream; b) from the social media data stream, identifying at least one photographic image from a first social media channel and at least one photographic image from a second social media channel; c) comparing the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel via a plurality of image recognition steps, comprising for each image;
i) converting the image down to 8×
8 pixels to remove unrequired details and to produce a low frequency image with 64 total pixels;
ii) converting the 8×
8 pixels into grey scale values to simplify further computation;
iii) computing a mean value for all of the 8×
8 pixels;
iv) creating a 64 bit hash where each bit is set to 0 or 1 depending on whether the grey scale value is above or below the mean value;
v) comparing two images using Hamming distance to produce a global image comparison output;d) comparing the global image comparison output to a threshold to produce a set of resultant global images, wherein the set of resultant global images comprises at least one pair of photographic images from the first social media channel and the second social media channel of the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel having a global image comparison output less than the threshold; e) comparing the set of resultant global images from the first social media channel and the set of resultant global images from the second social media channel via a plurality of image recognition steps, comprising for each image, i) identifying one or more features/keypoints;
ii) assigning descriptors to the one or more features/keypoints so that they can be recognized and compared to other images;
iii) weighing the one or more features/keypoints;
iv) matching the one or more features/keypoints;f) from the social media data stream, i) identifying a plurality of geolocation data inputs related to the person and/or social media connections of the person;
ii) performing a centroid-based analysis on the plurality of geolocation data inputs to form clusters;
iii) classifying and filtering the clusters; and
iv) selecting one or more probable locations of the person based upon the clusters; andg) determining and corroborating with confidence the photographic images and the geolocation data inputs which are assignable to the person and creating the integrated, collated social media profile, displayable on a dashboard, for the person, wherein the integrated, collated social media profile for the person comprises at least one matched photographic image of the set of resultant global images. - View Dependent Claims (18)
-
-
19. A non-transitory computer readable medium having stored thereon computer-executable instructions, the computer executable instructions causing a processor to execute a method of creating an integrated, collated social media profile for a person which comprises:
-
a) probing social media outlets and channels using one or more of social media application programming interfaces (APIs), data provided by third party data compilers, data aggregated directly from web crawlers, using selected person-specific search parameters and person specific context data to form a social media data stream; b) from the social media data stream, identifying at least one photographic image from a first social media channel and at least one photographic image from a second social media channel; c) comparing the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel via a plurality of image recognition steps, comprising for each image;
i) converting the image down to 32×
32 pixels in order to simplify Discrete Cosine Transform (DCT) computation;
ii) computing the DCT of the 32×
32 image to produce a 32×
32 collection of frequency scalars;
iii) selecting the top left 8×
8 scalars and disregarding remaining scalars;
iv) computing a mean average of the selected DCT scalars;
v) creating a 64 bit hash where each bit is set to 0 or 1 depending on whether each of the 64 DCT scalars is above or below the mean average;
vi) comparing two images using Hamming distance to produce a global image comparison output;d) comparing the global image comparison output to a threshold to produce a set of resultant global images, wherein the set of resultant global images comprises at least one pair of photographic images from the first social media channel and the second social media channel of the at least one photographic image from the first social media channel and the at least one photographic image from the second social media channel having a global image comparison output less than the threshold; e) comparing the set of resultant global images from the first social media channel and the set of resultant global images from the second social media channel via a plurality of image recognition steps, comprising for each image, i) identifying one or more features/keypoints;
ii) assigning descriptors to the one or more features/keypoints so that they can be recognized and compared to other images;
iii) weighing the one or more features/keypoints;
iv) matching the one or more features/keypoints;f) from the social media data stream, i) identifying a plurality of geolocation data inputs related to the person and/or social media connections of the person;
ii) performing a centroid-based analysis on the plurality of geolocation data inputs to form clusters;
iii) classifying and filtering the clusters; and
iv) selecting one or more probable locations of the person based upon the clusters; andg) determining and corroborating with confidence the photographic images and the geolocation data inputs which are assignable to the person and creating the integrated, collated social media profile, displayable on a dashboard, for the person, wherein the integrated, collated social media profile for the person comprises at least one matched photographic image of the set of resultant global images. - View Dependent Claims (20)
-
Specification