SYSTEMS AND METHODS FOR SCENE CHANGE RECOMMENDATIONS
Systems and methods are disclosed for indicating whether a scene in a media asset corresponds to a scene for which a scene change was previously requested. A media player client monitors scene sequences in a media asset to anticipate whether the user will initiate a scene change request based on historical scene change data. In response to detecting a scene sequence for which the user is expected to request a scene change, the media player client determines whether a scene change is necessary and outputs an indication of whether a scene change is recommended.
1 - 50 . (canceled)
51 . A method comprising:
training, based on a plurality of training data structures of a user profile, a machine learning model to classify a scene, wherein the machine learning model is configured to:
receive an input data structure; and
output an output binary classifier indicating whether a scene change is recommended;
receiving a first scene change request for a first scene of a media asset;
generating a first data structure, wherein the first data structure indicates metadata for the first scene;
inputting the first data structure into the machine learning model; and
receiving a binary operator indicating that the first scene should not be skipped.
52 . The method of claim 51 wherein the training the machine learning model to classify the scene comprises:
identifying a plurality of scene change requests in a viewing history associated with the user profile; and
for each respective scene change request of the plurality of scene change requests:
determining a respective scene category of the respective scene associated with the scene change request; and
generating a respective training data structure of the plurality of training data structures, wherein the respective training data structure comprises the respective scene category of the respective scene associated with the scene change request and the respective scene change request.
53 . The method of claim 52 wherein the training the machine learning model to classify the scene further comprises:
identifying a plurality of unskipped scenes in a viewing history associated with the user profile; and
for each respective unskipped scenes of the plurality of unskipped scenes:
determining a respective scene category of the respective unskipped scene; and
generating a respective training data structure of the plurality of training data structures, wherein the respective training data structure comprises the respective scene category of the respective unskipped scene and an indicator that a scene change request was not received.
54 . The method of claim 53 wherein the outputting the output binary classifier indicating whether a scene change is recommended comprises:
determining that the input data structure comprises a first scene category;
determining, by the machine learning model, that the first scene category corresponds to a scene category associated with a scene change request; and
wherein the binary operator indicates that the scene change is recommended.
55 . The method of claim 53 wherein the outputting the output binary classifier indicating whether a scene change is recommended comprises:
determining that the input data structure comprises a first scene category;
determining, by the machine learning model, that the first scene category corresponds to a scene category associated with at least one of the unskipped scenes of the plurality of unskipped scenes; and
wherein the binary operator indicates that the scene change is not recommended.
56 . The method of claim 51 wherein the training the machine learning model to classify the scene comprises:
identifying a scene change request associated with a second scene in a viewing history associated with the user profile;
determining a sequence of scenes preceding the second scene; and
determining a plurality of scene categories associated with the sequence of scenes preceding the second scene; and
generating a respective training data structure of the plurality of training data structures, wherein the respective training data structure comprises the plurality of scene categories associated with the sequence of scenes preceding the second scene and the respective scene change request.
57 . The method of claim 51 further comprising generating for display a warning that the scene change is not recommended.
58 . The method of claim 51 wherein the generating the first data structure comprises:
determining, based on metadata of the media asset, a plurality of categories for a sequence of scenes preceding the first scene.
59 . The method of claim 51 further comprising:
in response to receiving the binary operator indicating that the first scene should not be skipped, executing a modified scene change request.
60 . The method of claim 59 , wherein the first scene change request comprises a fast-forwarding request, wherein executing the modified scene change request comprises:
determining a number of frames skipped in a period of time for a fast-forward request; and
modifying the fast-forwarding request by reducing the number of frames skipped in the period of time.
61 . A system comprising:
control circuitry configured to:
train, based on a plurality of training data structures of a user profile, a machine learning model to classify a scene, wherein the machine learning model is configured to:
receive an input data structure; and
output an output binary classifier indicating whether a scene change is recommended;
input/output circuitry configured to:
receive a first scene change request for a first scene of a media asset;
the control circuitry further configured to:
generate a first data structure, wherein the first data structure indicates metadata for the first scene;
input the first data structure into the machine learning model; and
receive a binary operator indicating that the first scene should not be skipped.
62 . The system of claim 61 wherein the control circuitry is configured to train the machine learning model to classify the scene by:
identifying a plurality of scene change requests in a viewing history associated with the user profile; and
for each respective scene change request of the plurality of scene change requests:
determining a respective scene category of the respective scene associated with the scene change request; and
generating a respective training data structure of the plurality of training data structures, wherein the respective training data structure comprises the respective scene category of the respective scene associated with the scene change request and the respective scene change request.
63 . The system of claim 62 wherein the control circuitry is further configured to train the machine learning model to classify the scene by:
identifying a plurality of unskipped scenes in a viewing history associated with the user profile; and
for each respective unskipped scenes of the plurality of unskipped scenes:
determining a respective scene category of the respective unskipped scene; and
generating a respective training data structure of the plurality of training data structures, wherein the respective training data structure comprises the respective scene category of the respective unskipped scene and an indicator that a scene change request was not received.
64 . The system of claim 63 wherein the control circuitry is configured to output the output binary classifier indicating whether a scene change is recommended by:
determining that the input data structure comprises a first scene category;
determining, by the machine learning model, that the first scene category corresponds to a scene category associated with a scene change request; and
wherein the binary operator indicates that the scene change is recommended.
65 . The system of claim 63 wherein the control circuitry is configured to output the output binary classifier indicating whether a scene change is recommended by:
determining that the input data structure comprises a first scene category;
determining, by the machine learning model, that the first scene category corresponds to a scene category associated with at least one of the unskipped scenes of the plurality of unskipped scenes; and
wherein the binary operator indicates that the scene change is not recommended.
66 . The system of claim 61 wherein the control circuitry is configured to train the machine learning model to classify the scene by:
identifying a scene change request associated with a second scene in a viewing history associated with the user profile;
determining a sequence of scenes preceding the second scene; and
determining a plurality of scene categories associated with the sequence of scenes preceding the second scene; and
generating a respective training data structure of the plurality of training data structures, wherein the respective training data structure comprises the plurality of scene categories associated with the sequence of scenes preceding the second scene and the respective scene change request.
67 . The system of claim 61 wherein the control circuitry is further configured to generate for display a warning that the scene change is not recommended.
68 . The system of claim 61 wherein the control circuitry is configured to generate the first data structure by:
determining, based on metadata of the media asset, a plurality of categories for a sequence of scenes preceding the first scene.
69 . The system of claim 61 wherein the control circuitry is further configured to:
in response to receiving the binary operator indicating that the first scene should not be skipped, executing a modified scene change request.
70 . The system of claim 69 , wherein the first scene change request comprises a fast-forwarding request, and wherein the control circuitry is configured to execute the modified scene change request by:
determining a number of frames skipped in a period of time for a fast-forward request; and
modifying the fast-forwarding request by reducing the number of frames skipped in the period of time.