Face search in personals

• US 9,753,948 B2
• Filed: 05/27/2008
• Issued: 09/05/2017
• Est. Priority Date: 05/27/2008
• Status: Active Grant

• Current Assignee: Match Group, LLC (Match Group, Inc.)
• Original Assignee: Match.com LLC (Match Group, Inc.)
• Inventors: Lo, Kin Man
• Primary Examiner(s): Lu, Kuen
• Application Number:
  US12/127,577
• Publication Number:
  US 20090299961A1
• Time in Patent Office: 3,388 Days
• Field of Search: 707769, 707733, 382118
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: G06F 18/2135 based on approximation criteria, e.g. principal component analysis
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: G06V 10/7715 Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

A device, system and method to enable searching of personal profiles in the context of on-line dating that includes the ability to determine the personal profiles that have images that most closely resemble a target image.

Claims

1. A network device to manage communications over a network, comprising:

a transceiver to send and receive data over a network; and

a processor that is operative to enable actions for;

receiving a request to search a plurality of different personal profiles, each of the plurality of different personal profiles includes an image having a corresponding reduced image data set comprising a face space, wherein the request comprises a personal characteristic and a target image, and receiving the request comprises;

receiving, from a device associated with a personal profile from the plurality of different personal profiles, selection of the personal characteristic, and

receiving, from the device associated with the personal profile, selection of the target image, the target image depicting at least one feature that is desirable in a potential match for the personal profile, wherein the at least one feature depicted by the target image comprises eyes or body type of the potential match;

calculating a reduced image data set for the target image, the reduced image data set comprising a target image face space;

retrieving from the plurality of different personal profiles a subset of personal profiles that contain a match to the personal characteristic;

retrieving from the subset of personal profiles a subset of reduced image data sets;

determining correlations between the reduced image data set for the target image and each reduced image data set in the subset of reduced image data sets based on the at least one feature;

ranking the subset of reduced image data sets associated with the personal characteristic according to the correlations; and

providing at least one personal profile from the plurality of different personal profiles in an order determined based on the ranking of the subset of reduced image data sets.

2. The network device of claim 1, wherein the reduced image data set for an image constitutes a reduced set of data that retains principal characteristics of the image.

3. The network device of claim 1, wherein the step of calculating the reduced image data set for the target image further comprises:

calculating a face space multiplier;

processing the target image to produce a mean subtracted, normalized target image; and

calculating the reduced image data set for the target image by multiplying the face space multiplier by the mean subtracted, normalized target image.

4. The network device of claim 3, wherein the face space multiplier contains a number of eigenvectors.

5. The network device of claim 4, wherein the number of eigenvectors in the face space multiplier is substantially less than the total possible number of eigenvectors.

6. The network device of claim 3, wherein the face space multiplier is calculated prior to receiving the request to search personal profiles.

7. The network device of claim 1, wherein the target image is received based on a selection of at least one of a set of generic features that mimic an image of an idealized person.

8. The network device of claim 1, wherein:

the target image has a number of data points;

the reduced image data set for the target image has a number of data points which is less than the number of data points contained in the target image; and

the number of data points of the reduced image data set for the target image used to perform the determining and ranking is dependent on availability of system resources for performing the determining and ranking;

wherein the number of data points of the reduced image data is dynamically varied based on the availability of the system resources.

9. A method for managing a personal profile search over a network, comprising:

receiving a request to search a plurality of different personal profiles, each of the plurality of different personal profiles includes an image having a corresponding reduced image data set comprising a face space, wherein the request comprises a personal characteristic and a target image, and receiving the request comprises;

receiving, from a device associated with a personal profile from the plurality of different personal profiles, selection of the personal characteristic, and

receiving, from the device associated with the personal profile, selection of the target image, the target image depicting at least one feature that is desirable in a potential match for the personal profile, wherein the at least one feature depicted by the target image comprises eyes or body type of the potential match;

calculating a reduced image data set for the target image, comprising a target image face space;

retrieving from the plurality of different personal profiles a subset of personal profiles that contain a match to the personal characteristic;

retrieving from the subset of personal profiles a subset of reduced image data sets;

determining correlations between the reduced image data set for the target image and each reduced image data set in the subset of reduced image data sets based on the at least one or more feature;

ranking the subset of reduced image data sets associated with the personal characteristic based on the correlations; and

providing at least one personal profile from the plurality of different personal profiles in an order determined based on the ranking of the subset of reduced image data sets.

10. The method of claim 9, wherein:

the target image has a number of data points;

the reduced image data set for the target image has a number of data points which is less than the number of data points contained in the target image; and

the number of data points of the reduced image data set for the target image used to perform the determining and ranking is dependent on availability of system resources for performing the determining and ranking.

11. The method of claim 10, wherein:

the number of data points of the reduced image data set is made dependent on availability of system resources by dynamically varying the dimension of the reduced image data set used to calculate the determining and ranking according to the availability of system resources for performing the determining and ranking.

12. The method of claim 9, wherein the target image is received based on a selection of at least one of a set of generic features that mimic an image of an idealized person.

13. A system for managing a personal profile search over a network, comprising:

a plurality of client devices, wherein each client device is configured to receive and send messages over the network; and

a network device that is in communication with the plurality of client devices, and is configured and arranged to perform actions, including;

receiving a request for a personal profile search within a plurality of different personal profiles from one client device in the plurality of client devices, the one client device being associated with a personal profile from the plurality of different personal profiles, and each one of the plurality of different personal profiles includes a digitized image, wherein the request comprises a personal characteristic and a target image, and receiving the request comprises;

receiving, from the one client device, selection of the personal characteristic, and

receiving, from the one client device, selection of the target image, the target image depicting at least one feature that is desirable in a potential match for the personal profile, wherein the at least one feature depicted by the target image comprises eyes or body type of the potential match;

calculating a reduced image data set for the target image comprising a target image face space;

querying a personal profile database having the plurality of different personal profiles for a subset of personal profiles that contain a match to the personal characteristic;

comparing the reduced image data set for the target image against digitized images associated with the subset of personal profiles that contain the match to the personal characteristic to determine which of the digitized images most closely resemble the target image based on the at least one desired feature;

sorting the digitized images based on the resemblance to the target image;

providing to the client device at least one personal profile from the plurality of different profiles in an order determined based on the ranking of the digitized images.

14. The system of claim 13, further comprising:

calculating percentages of correlation between each of the digitized images and the target image; and

transmitting the percentages of correlation to the client device.

15. The system of claim 13, wherein the target image is received based on a selection of at least one of a set of generic features that mimic an image of an idealized person.

16. The system of claim 13, wherein:

the target image has a number of data points;

the reduced image data set for the target image has a number of data points which is less than the number of data points contained in the target image; and

the number of data points of the reduced image data set for the target image used to perform the determining and ranking is dependent on availability of system resources for performing the determining and ranking;

wherein the number of data points of the reduced image data is dynamically varied based on the availability of the system resources.

17. A machine-readable tangible and non-transitory medium having information recorded thereon for managing a personal profile search over a network, wherein the information, when read by the machine, causes the machine to perform the following:

receiving a request to search a plurality of different personal profiles, each of the plurality of different personal profiles including an image having a corresponding reduced image data set comprising a face space, wherein the request comprises a personal characteristic and a target image, and receiving the request comprises;

receiving, from a device associated with a personal profile from the plurality of different personal profiles, selection of the personal characteristic, and

receiving, from the device associated with the personal profile, selection of the target image, the target image depicting at least one feature that is desirable in a potential match for the personal profile, wherein the at least one feature depicted by the target image comprises eyes or body type of the potential match;

calculating a reduce image data set for the target image comprising a target image face space;

retrieving from the plurality of different personal profiles a subset of personal profiles that contain a match to the personal characteristic;

retrieving from the subset of personal profiles a subset of reduced image data sets;

determining correlations between the reduced image data set for the target image and each reduced image data set in the subset of reduced image data sets based on the at least one desired feature;

ranking the subset of reduced image data sets associated with the personal characteristic based on the correlations; and

providing at least one personal profile from the plurality of different personal profiles in an order determined based on the ranking of the subset of reduced image data sets.

18. The machine-readable tangible and non-transitory medium of claim 17, wherein the target image is received based on a selection of at least one of a set of generic features that mimic an image of an idealized person.

19. The machine-readable tangible and non-transitory medium of claim 17, wherein:

the target image has a number of data points;

the reduced image data set for the target image has a number of data points which is less than the number of data points contained in the target image; and

the number of data points of the reduced image data set for the target image used to perform the determining and ranking is dependent on availability of system resources for performing the determining and ranking;

wherein the number of data points of the reduced image data is dynamically varied based on the availability of the system resources.