Machine learning model-based detection of content type

Patent metadata

Field	Value
Publication number	US-12394097-B2
Application number	US-202318098935-A
Country	US
Kind code	B2
Filing date	Jan 19, 2023
Priority date	Jan 19, 2023
Publication date	Aug 19, 2025
Grant date	Aug 19, 2025

Abstract

A system includes a hardware processor, and a memory storing a software code and at least one machine learning (ML) model trained to distinguish between a plurality of content types. The hardware processor executes the software code to receive a content file including data identifying a dataset contained by the content file as being a first content type of the plurality of content types; predict, using the at least one ML model and the dataset, based on at least one image parameter, a first probability that a content type of the dataset matches the first content type identified by the data; and determine, based on the first probability, that the content type of the dataset (i) is the first content type identified by the data, (ii) is not the first content type identified by the data, or (iii) is of an indeterminate content type.

First claim

What is claimed is:

1. A system comprising:
a hardware processor;
a system memory storing a software code and at least one machine learning (ML) model trained to distinguish between a plurality of content types;
the hardware processor configured to execute the software code to:
receive a content file including data identifying a dataset contained by the content file as being a first content type of the plurality of content types;
predict, using the at least one ML model and the dataset, based on at least one image parameter, a first probability that a content type of the dataset matches the first content type identified by the data; and
determine, based on the first probability, that the content type of the dataset (i) is the first content type identified by the data, (ii) is not the first content type identified by the data, or (iii) is of an indeterminate content type.

2. The system of claim 1 , wherein when determining determines that the content type of the dataset is the first content type identified by the data, the hardware processor is further configured to execute the software code to:
output the content file to a content processing system or a content distribution system in an automated process.

3. The system of claim 1 , wherein when determining determines that the content type of the dataset is not the first content type identified by the data or is of the indeterminate content type, the hardware processor is further configured to execute the software code to:
flag the content file for human review.

4. The system of claim 1 , wherein the hardware processor is further configured to execute the software code to:
predict, using the at least one ML model and the dataset, and based on the at least one image parameter, a second probability that the content type of the dataset matches a second content type of the plurality of content types;
wherein determining that the content type of the dataset (i) is the first content type identified by the data, (ii) is not the first content type identified by the data, or (iii) is of the indeterminate content type is further based on the second probability.

5. The system of claim 1 , wherein the at least one image parameter comprises an electro-optical transfer function (EOTF) of the dataset.

6. The system of claim 1 , wherein the at least one image parameter comprises one or more of an electro-optical transfer function (EOTF), a quantization range, or a color encoding primary of the dataset.

7. The system of claim 1 , wherein the plurality of content types comprise standard dynamic range (SDR) content and high dynamic range (HDR) content.

8. The system of claim 1 , wherein the at least one ML model comprises a random decision forest or a neural network.

9. A method for use by a system including a hardware processor and a system memory storing a software code and at least one machine learning (ML) model trained to distinguish between a plurality of content types, the method comprising:
receiving, by the software code executed by the hardware processor, a content file including data identifying a dataset contained by the content file as being a first content type of the plurality of content types;
predicting, by the software code executed by the hardware processor using at least one MIL model and the dataset, based on at least one image parameter, a first probability that a content type of the dataset matches the first content type identified by the received data; and
determining, by the software code executed by the hardware processor based on the first probability, that the content type of the dataset (i) is the first content type identified by the data, (ii) is not the first content type identified by the data, or (iii) is of an indeterminate content type.

10. The method of claim 9 , wherein when determining determines that the content type of the dataset is the first content type identified by the data, the method further comprises:
outputting, by the software code executed by the hardware processor, the content file to a content processing system or a content distribution system in an automated process.

11. The method of claim 9 , wherein when determining determines that the content type of the dataset is not the first content type identified by the data or is of the indeterminate content type, the method further comprises:
flagging, by the software code executed by the hardware processor, the content file for human review.

12. The method of claim 9 , further comprising:
predicting, by the software code executed by the hardware processor using the at least one ML model and the dataset, and based on the at least one image parameter, a second probability that a content type of the dataset matches a second content type of the plurality of content types;
wherein determining that the content type of the dataset (i) is the first content type identified by the data, (ii) is not the first content type identified by the data, or (iii) is of the indeterminate content type is further based on the second probability.

13. The method of claim 9 , wherein the at least one image parameter comprises an electro-optical transfer function (EOTF) of the dataset.

14. The method of claim 9 , wherein the at least one image parameter comprises one or more of an electro-optical transfer function (EOTF), a quantization range, or color encoding primaries of the dataset.

15. The method of claim 9 , wherein the plurality of content types comprise standard dynamic range (SDR) content and high dynamic range (HDR) content.

16. The method of claim 9 , wherein the at least one ML model comprises one of a random decision forest or a neural network.

17. A method for training a machine learning (ML) model to distinguish between a plurality of content types, the method comprising:
obtaining a plurality of image datasets;
generating, for each of the plurality of image datasets, a respective statistical representation of each of one or more variables for use in detecting an image parameter, to provide a plurality of statistical representations;
correlating each of the plurality of statistical representations with one of the plurality of content types; and
training the ML model, using the training data, to predict a first probability that a content type of another dataset matches at least one of the plurality of content types.

18. The method of claim 17 , wherein the image parameter comprises an electro-optical transfer function (EOTF), a quantization range, or a color encoding primary of an image.

19. The method of claim 17 , wherein the plurality of content types comprise standard dynamic range (SDR) content and high dynamic range (HDR) content.

20. The method of claim 17 , wherein the ML model comprises one of a random decision forest or a neural network.

Assignees

• Disney Entpr Inc

Inventors

• Jacobs Mitchel
• Chakrahari Anil
• Kuang Ellie
• Smith Michael

Classifications

• G06T2207/20081

  Training; Learning · CPC title

• G06T2207/20208

  High dynamic range [HDR] image processing · CPC title

• H04N21/23439

  for generating different versions · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06T7/90 Primary

  Determination of colour characteristics · CPC title