Granted Patent B2

US 12,374,017 · App. 18/388,987 · Granted Jul 29, 2025

Inventors: Kai Zhou (Wiener Neudorf, AT); Kenneth Au (Santa Monica, CA)

Assignee: Snap Inc.

G06T13/40G06F16/90335G06F18/21G06T19/006G06V20/30H04L67/535

Abstract

Claims (97)

1. A method comprising:

accessing an animated image file;

identifying a plurality of elements of the animated image file by applying a plurality of machine-learning programs, the plurality of machine-learning programs comprising a visual environment machine learning application, a human model machine learning application, and a camera trajectory machine learning application that trains an augmented reality effect machine learning model, wherein a camera trajectory machine learning model of the camera trajectory machine learning application is trained to identify a movement of a camera resulting in different angles in the animated image file;

generating, using a variant element combination machine-learning application, a variation of one or more elements of the plurality of elements; and

generating a variant animated image file by applying the variation of the one or more elements to the animated image file.

2. The method of claim 1 , further comprising:

accessing messages from a communication platform;

identifying a plurality of animated image files in the messages;

measuring a trending metric for each animated image file based on a number of times each animated image file is shared on the communication platform within a predefined period of time; and

identifying the animated image file having a highest trending metric within the predefined period of time.

3. The method of claim 1 , wherein the plurality of machine-learning programs comprises an animation machine learning application, and a text machine learning application.

4. The method of claim 3 , further comprising:

parsing the animated image file by applying one of the visual environment machine learning application, the human model machine learning application, the camera trajectory machine learning application, the animation machine learning application, the augmented reality effect machine learning application, and the text machine learning application to the animated image file to identify the plurality of elements displayed in the animated image file.

5. The method of claim 1 , wherein one of the plurality of elements comprises a visual environment element, a human model element, a camera trajectory element, an animation element, an augmented reality effect element, or a text element.

6. The method of claim 1 , further comprising:

identifying a first element displayed in the animated image file; and

accessing a first variation of the first element from an element library corresponding to the first element,

wherein applying the variation of the one or more elements comprises:

applying the first variation to the first element of the animated image file,

wherein generating the variant animated image file is based on the first variation of the first element of the animated image file.

7. The method of claim 1 , further comprising:

training a visual environment machine learning model with the visual environment machine learning application;

training a human model machine learning model with the human model machine learning application;

training a camera trajectory machine learning model with the camera trajectory machine learning application;

training an animation machine learning model with an animation machine learning application;

training an augmented reality effect machine learning model with the augmented reality effect machine learning application;

training a text machine learning model with a text machine learning application; and

training a variant element combination machine learning model with a variant element combination machine learning application, the variant element combination machine learning model being trained to identify combinations of elements variations of top trending variants.

8. The method of claim 1 , further comprising:

publishing the variant animated image file on a communication platform;

measuring a trending metric of the variant animated image file based on a number of times the variant animated image file is shared on the communication platform within a predefined period of time;

accessing the trending metric of each variant animated image file on the communication platform;

identifying a top variant animated image file based on top trending metrics;

identifying a variation of elements corresponding to the top variant animated image file;

detecting an animated image file sharing element at a device operating an application of the communication platform; and

generating, at the device, a recommendation dialog box that recommends the variation of elements.

9. The method of claim 1 , further comprising:

applying a combination of variations of the one or more elements to the animated image file;

generating a plurality of variant animated image files based on the combination of variations;

publishing the plurality of variant animated image files on a communication platform;

measuring a trending metric of each published variant animated image file based on a number of times each published variant animated image file is shared on the communication platform within a predefined period of time;

identifying a published variant animated image file with a highest trending metric from the plurality of variant animated image files; and

identifying the combination of variations of the one or more elements of the identified published variant animated image file.

10. The method of claim 9 , further comprising:

accessing a first animated image file;

applying a combination of variations of the one or more elements of the identified published variant animated image file to the first animated image file; and

generating a second animated image file based on the applied combination of variations of the one or more elements of the identified published variant animated image file to the second animated image file, wherein the first animated image file comprises a first animated GIF file.

11. A computing apparatus comprising:

a processor; and

a memory storing instructions that, when executed by the processor, configure the apparatus to perform operations comprising:

accessing an animated image file;

identifying a plurality of elements of the animated image file by applying a plurality of machine-learning programs, the plurality of machine-learning programs comprising a visual environment machine learning application, a human model machine learning application, and a camera trajectory machine learning application that trains an augmented reality effect machine learning model, wherein a camera trajectory machine learning model of the camera trajectory machine learning application is trained to identify a movement of a camera resulting in different angles in the animated image file;

generating, using a variant element combination machine-learning application, a variation of one or more elements of the plurality of elements; and

generating a variant animated image file by applying the variation of the one or more elements to the animated image file.

12. The computing apparatus of claim 11 , wherein the operations further comprise:

accessing messages from a communication platform;

identifying a plurality of animated image files in the messages;

measuring a trending metric for each animated image file based on a number of times each animated image file is shared on the communication platform within a predefined period of time; and

identifying the animated image file having a highest trending metric within the predefined period of time.

13. The computing apparatus of claim 11 , wherein the plurality of machine-learning programs comprises an animation machine learning application, and a text machine learning application.

14. The computing apparatus of claim 13 , wherein the operations further comprise:

parsing the animated image file by applying one of the visual environment machine learning application, the human model machine learning application, the camera trajectory machine learning application, the animation machine learning application, the augmented reality effect machine learning application, and the text machine learning application to the animated image file to identify the plurality of elements displayed in the animated image file.

15. The computing apparatus of claim 11 , wherein one of the plurality of elements comprises a visual environment element, a human model element, a camera trajectory element, an animation element, an augmented reality effect element, or a text element.

16. The computing apparatus of claim 11 , wherein the operations further comprise:

identifying a first element displayed in the animated image file; and

accessing a first variation of the first element from an element library corresponding to the first element,

wherein applying the variation of the one or more elements comprises:

applying the first variation to the first element of the animated image file,

wherein generating the variant animated image file is based on the first variation of the first element of the animated image file.

17. The computing apparatus of claim 11 , wherein the operations further comprise:

training a visual environment machine learning model with the visual environment machine learning application;

training a human model machine learning model with the human model machine learning application;

training a camera trajectory machine learning model with the camera trajectory machine learning application;

training an animation machine learning model with an animation machine learning application;

training an augmented reality effect machine learning model with the augmented reality effect machine learning application;

training a text machine learning model with a text machine learning application; and

training a variant element combination machine learning model with a variant element combination machine learning application, the variant element combination machine learning model being trained to identify combinations of elements variations of top trending variants.

18. The computing apparatus of claim 11 , wherein the operations further comprise:

publishing the variant animated image file on a communication platform;

measuring a trending metric of the variant animated image file based on a number of times the variant animated image file is shared on the communication platform within a predefined period of time;

accessing the trending metric of each variant animated image file on the communication platform;

identifying a top variant animated image file based on top trending metrics;

identifying a variation of elements corresponding to the top variant animated image file;

detecting an animated image file sharing element at a device operating an application of the communication platform; and

generating, at the device, a recommendation dialog box that recommends the variation of elements.

19. The computing apparatus of claim 11 , wherein the operations further comprise:

applying a combination of variations of the one or more elements to the animated image file;

generating a plurality of variant animated image files based on the combination of variations;

publishing the plurality of variant animated image files on a communication platform;

measuring a trending metric of each published variant animated image file based on a number of times each published variant animated image file is shared on the communication platform within a predefined period of time;

identifying a published variant animated image file with a highest trending metric from the plurality of variant animated image files; and

identifying the combination of variations of the one or more elements of the identified published variant animated image file.

20. A non-transitory computer-readable storage medium, the computer-readable storage medium including instructions that when executed by a computer, cause the computer to perform operations comprising:

accessing an animated image file;

identifying a plurality of elements of the animated image file by applying a plurality of machine-learning programs, the plurality of machine-learning programs comprising a visual environment machine learning application, a human model machine learning application, and a camera trajectory machine learning application that trains an augmented reality effect machine learning model, wherein a camera trajectory machine learning model of the camera trajectory machine learning application is trained to identify a movement of a camera resulting in different angles in the animated image file;

generating, using a variant element combination machine-learning application, a variation of one or more elements of the plurality of elements; and

generating a variant animated image file by applying the variation of the one or more elements to the animated image file.

Assignments (1)

Continuity (2)

References Cited (32)