Granted Patent B2

US 10,592,074 · App. 16/403,336 · Granted Mar 17, 2020

Inventors: Jehan Hamedi (South Boston, MA); Zachary McDonald Halloran (Braintree, MA)

Assignee: Adhark, Inc.

G06F3/0482G06F16/334G06F17/2705G06K9/00684G06K9/6215G06N20/00

Abstract

Claims (108)

1. A system for analyzing visual content items, the system comprising:

a memory; and

one or more processors coupled to the memory, wherein the one or more processors include programmed instructions to:

access, from a mobile computing device, a plurality of visual content items captured by the mobile computing device;

evaluate the plurality of visual content items to determine a respective similarity score between the plurality of visual content items and each of a plurality of identity categories, wherein evaluating the plurality of visual content items comprises applying at least one machine learning model to the plurality of visual content items for each of the plurality of identity categories;

determine a subset of the identity categories for which the respective similarity scores exceed a similarity threshold;

generate information corresponding to a first graphical user interface (GUI), the first GUI comprising a visual representation of each identity category of the subset of identity categories and a user-selectable interface element for each identity category of the subset of identity categories;

provide the information corresponding to the first GUI to the mobile computing device to cause the mobile computing device to render the first GUI via an electronic display of the mobile device;

receive a first user input corresponding to a selection of at least one identity category of the subset of identity categories;

identify a target audience for a user of the mobile computing device;

determine activity data for a plurality of published content items retrieved from at least one content source, the published content items selected based on the at least one identity category, wherein the activity data comprises data relating to engagement of viewers with the plurality of published content items;

train a second machine learning model to determine a predicted performance score based on the published content items and the activity data;

evaluate each visual content item of the plurality of visual content items using the second machine learning model to generate a respective predicted performance score for each visual content item of the plurality of visual content items;

generate information corresponding to a second GUI, the second GUI comprising a visual representation of at least a subset of the plurality of visual content items and the respective predicted performance scores for the subset of the plurality of visual content items; and

provide the information corresponding to the second GUI to the mobile computing device to cause the mobile computing device to render the second GUI via the electronic display of the mobile device.

2. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

provide information corresponding to a third GUI to the mobile computing device, the third GUI comprising a list of a plurality of candidate audiences;

receive, from the mobile computing device, a second user input corresponding to a selection of a first candidate audience of the plurality of candidate audiences; and

identify the target audience based on the second user input.

3. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

access a first account on the target platform corresponding to the user of the mobile computing device;

identify a plurality of second accounts on the target platform, each second account linked with the first account; and

identify the target audience based on the plurality of second accounts.

4. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

detect that a new visual content item has been captured by the mobile computing device, the new visual content item not included in the plurality of visual content items; and

automatically evaluate the new visual content item using the second machine learning model based on the published content items and the activity data to generate a respective predicted performance score for the new visual content item, responsive to detecting that the new visual content item has been capture by the mobile computing device.

5. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

generate the information corresponding to the second GUI such that the second GUI comprises the visual representations of the subset of the visual content items arranged in a rectangular array, wherein the respective predicted performance scores overlap at least a portion of the visual representations of their respective visual content items within the second GUI.

6. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

access a new visual content item from a content source remote from the mobile computing device;

evaluate the new visual content item using the second machine learning model based on the published content items and the activity data to generate a respective predicted performance score for the new visual content item;

determine that the predicted performance score for the new visual content item exceeds a predetermined predicted performance threshold; and

provide the new visual content item to the mobile computing device.

7. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

receive, from the mobile computing device, a second user input specifying a time period;

determine a timestamp of each visual content item of the plurality of visual content items; and

identify the subset of the plurality of visual content items for the second GUI such that the respective timestamp of each visual content item of the subset of visual content items falls within the time period.

8. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

identify a target content item characteristic;

for each visual content item of the plurality of visual content items, evaluate the visual content item to determine whether the visual content item includes the target content item characteristic; and

identify the subset of the plurality of visual content items for the second GUI such that each visual content item of the subset of visual content items includes the target content item characteristic.

9. The system of claim 8 , wherein the one or more processors further include programmed instructions to:

receive, from the mobile computing device, a second user input corresponding to a search criteria text string; and

parse the search criteria text string to identify the target content item characteristic.

10. The system of claim 8 , wherein the one or more processors further include programmed instructions to:

receive, from the mobile computing device, a second user input corresponding to an audio signal; and

parse the audio signal to identify the target content item characteristic.

11. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

identify a target content platform; and

select the at least one content source from which the published content items are retrieved to correspond to the target content platform.

12. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

train the second machine learning model based in part on the target audience.

13. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

train the second machine learning model using one or more features extracted from image data corresponding to at least one of the published content items.

14. The system of claim 1 , wherein the one or more processors further include programmed instructions to:

evaluate the plurality of visual content items to determine the respective similarity score between the plurality of visual content items and each of the plurality of identity categories based on at least one of a percentage of the plurality of visual content items that correspond to each identity category or a quality rating of at least one of the plurality of visual content items that corresponds to each identity category.

15. A method for analyzing visual content items, the method comprising:

accessing, from a mobile computing device, a plurality of visual content items captured by the mobile computing device;

evaluating the plurality of visual content items to determine a respective similarity score between the plurality of visual content items and each of a plurality of identity categories, wherein evaluating the plurality of visual content items comprises applying at least one machine learning model to the plurality of visual content items for each of the plurality of identity categories;

determining a subset of the identity categories for which the respective similarity scores exceed a similarity threshold;

generating information corresponding to a first graphical user interface (GUI), the first GUI comprising a visual representation of each identity category of the subset of identity categories and a user-selectable interface element for each identity category of the subset of identity categories;

providing the information corresponding to the first GUI to the mobile computing device to cause the mobile computing device to render the first GUI via an electronic display of the mobile device;

receiving a first user input corresponding to a selection of at least one identity category of the subset of identity categories;

identifying a target audience for a user of the mobile computing device;

determining activity data for a plurality of published content items retrieved from at least one content source, the published content items selected based on the at least one identity category, wherein the activity data comprises data relating to engagement of viewers with the plurality of published content items;

training a second machine learning model to determine a predicted performance score based on the published content items and the activity data;

evaluating each visual content item of the plurality of visual content items using the second machine learning model to generate a respective predicted performance score for each visual content item of the plurality of visual content items;

generating information corresponding to a second GUI, the second GUI comprising a visual representation of at least a subset of the plurality of visual content items and the respective predicted performance scores for the subset of the plurality of visual content items; and

providing the information corresponding to the second GUI to the mobile computing device to cause the mobile computing device to render the second GUI via the electronic display of the mobile device.

16. The method of claim 15 , further comprising:

providing information corresponding to a third GUI to the mobile computing device, the third GUI comprising a list of a plurality of candidate audiences;

receiving, from the mobile computing device, a second user input corresponding to a selection of a first candidate audience of the plurality of candidate audiences; and

identifying the target audience based on the second user input.

17. The method of claim 15 , further comprising:

accessing a first account on the target platform corresponding to the user of the mobile computing device;

identifying a plurality of second accounts on the target platform, each second account linked with the first account; and

identifying the target audience based on the plurality of second accounts.

18. The method of claim 15 , further comprising:

detecting that a new visual content item has been captured by the mobile computing device, the new visual content item not included in the plurality of visual content items; and

evaluating the new visual content item using the second machine learning model based on the published content items and the activity data to generate a respective predicted performance score for the new visual content item, responsive to detecting that the new visual content item has been capture by the mobile computing device.

19. The method of claim 15 , further comprising:

generating the information corresponding to the second GUI such that the second GUI comprises the visual representations of the subset of the visual content items arranged in a rectangular array, wherein the respective predicted performance scores overlap at least a portion of the visual representations of their respective visual content items within the second GUI.

20. The method of claim 15 , further comprising:

accessing a new visual content item from a content source remote from the mobile computing device;

evaluating the new visual content item using the second machine learning model based on the published content items and the activity data to generate a respective predicted performance score for the new visual content item;

determining that the predicted performance score for the new visual content item exceeds a predetermined predicted performance threshold; and

providing the new visual content item to the mobile computing device.

21. The method of claim 15 , further comprising:

receiving, from the mobile computing device, a second user input specifying a time period;

determining a timestamp of each visual content item of the plurality of visual content items; and

identifying the subset of the plurality of visual content items for the second GUI such that the respective timestamp of each visual content item of the subset of visual content items falls within the time period.

22. The method of claim 15 , further comprising:

identifying a target content item characteristic;

for each visual content item of the plurality of visual content items, evaluating the visual content item to determine whether the visual content item includes the target content item characteristic; and

identifying the subset of the plurality of visual content items for the second GUI such that each visual content item of the subset of visual content items includes the target content item characteristic.

23. The method of claim 22 , further comprising:

receiving, from the mobile computing device, a second user input corresponding to a search criteria text string; and

parsing the search criteria text string to identify the target content item characteristic.

24. The method of claim 22 , further comprising:

receiving, from the mobile computing device, a second user input corresponding to an audio signal; and

parsing the audio signal to identify the target content item characteristic.

25. The system of claim 15 , further comprising:

identifying a target content platform; and

selecting the at least one content source from which the published content items are retrieved to correspond to the target content platform.

26. The method of claim 15 , further comprising:

training the second machine learning model based in part on the target audience.

27. The method of claim 15 , further comprising train the second machine learning model using one or more features extracted from image data corresponding to at least one of the published content items.

28. The method of claim 15 , further comprising evaluating the plurality of visual content items to determine the respective similarity score between the plurality of visual content items and each of the plurality of identity categories based on at least one of a percentage of the plurality of visual content items that correspond to each identity category or a quality rating of at least one of the plurality of visual content items that corresponds to each identity category.

Assignments (3)

Continuity (3)

Cited By (11)

US 12,198,403 US 12,223,689 US 12,249,117 US 12,249,118 US 12,254,669 US 12,260,611 US 12,277,748 US 12,277,749 US 12,283,084 US 12,340,557 US 12,482,227