US 12,342,046 B2 · Systems and methods for deep recommendations using signature analysis

Abstract

Systems and methods are described herein for providing content item recommendations based on a video. Using feature vectors corresponding to at least one frame of a video (e.g., generated based on texture and shape intensity of a frame), a recommendation system improves content recommendation using analytic and quantitative characteristics derived from a frame of a content item rather than merely manually labeled bibliographic data (e.g., a genre or producer). The recommendation system may generate a feature vector based on a texture, a shape intensity (e.g., generated from a Generalized Hough Transform), and temporal data corresponding to at least one frame of a video. The feature vector is analyzed using a machine learning model (e.g., a neural network) to produce a machine learning model output. The recommendation system causes a recommended content item to be provided based on the machine learning model output.

Claims20

Dependent claims

A computer-implemented method, comprising:

determining a user interaction of a user profile with a first content item;
determining, based on the user interaction, the first content item is of interest to the user profile;
identifying a first scene of the first content item;
generating a shape intensity associated with the first scene based on a representation of angles and distances at which lines depicted in the first scene are located;
determining one or more scenes of a second content item that are similar to the first scene based on a correlation between the generated shape intensity and a shape intensity of the one or more scenes of the second content item;
and
generating for display to the user profile a recommendation of the second content item, based on determining that the one or more scenes of the second content item are similar to the first scene, wherein generating for display to the user profile the recommendation of the second content item comprises generating for display the one or more scenes of the second content item.

4 more elements

A computer-implemented system, comprising:

memory;
control circuitry configured to:
determine a user interaction of a user profile with a first content item;
determine, based on the user interaction, the first content item is of interest to the user profile;
identify a first scene of the first content item;
generate a shape intensity associated with the first scene based on a representation of angles and distances at which lines depicted in the first scene are located;
determine one or more scenes of a second content item that are similar to the first scene based on a correlation between the generated shape intensity and a shape intensity of the one or more scenes of the second content item;
and
generate for display to the user profile a recommendation of the second content item, based on determining that the one or more scenes of the second content item are similar to the first scene, wherein the control circuitry generates for display to the user profile the recommendation of the second content item by generating for display the one or more scenes of the second content item.

6 more elements

Dependencies are inferred from the claim text.

Systems and methods for deep recommendations using signature analysis

Life of this patent

Systems and methods for deep recommendations using signature analysis

Life of this patent

Abstract

Claims20