Sensitivity assessment for media production using artificial intelligence
An automatic flagging of sensitive portions of a digital dataset for media production includes receiving the digital dataset comprising at least one of audio data, video data, or audio-video data for producing at least one media program. A processor identifies sensitive portions of the digital dataset likely to be in one or more defined content classifications, based at least in part on comparing unclassified portions of the digital dataset with classified portions of the prior media production using an algorithm, and generates a plurality of sensitivity tags each signifying a sensitivity assessment for a corresponding one of the sensitive portions. The processor may save the plurality of sensitivity tags each correlated to its corresponding one of the sensitive portions in a computer memory for use by a media production or localization team.
1 . A computer-implemented method for determining a sensitive portion of a dataset for media production, the computer-implemented method comprising:
receiving, by one or more processors, a digital dataset corresponding to a video source file;
processing, by the one or more processors, the digital dataset into a plurality of frames;
annotating, by the one or more processors via a machine-learning model, at least one of the plurality of frames with censorship metadata, wherein the censorship metadata includes a sensitive event;
mapping, by the one or more processors, utilizing a multidimensional tensor, the annotated plurality of frames to a sensitivity score based on the censorship metadata, wherein the mapping includes mapping the sensitive event to the sensitivity score;
assigning, by the one or more processors, a sensitivity tag to the sensitivity score, wherein the sensitivity tag indicates a censorship likelihood for the sensitive event;
updating, by the one or more processors, the multidimensional tensor to include the sensitivity tag, the multidimensional tensor having a plurality of dimensions including a sensitivity tag category, a localization region, and a sensitivity level; and
outputting, by the one or more processors, a visual representation of the sensitive event and the sensitivity tag to a display.
2 . The computer-implemented method of claim 1 , the computer-implemented method further comprising:
receiving, by the one or more processors, user feedback indicating a sensitivity score accuracy for the sensitivity score.
3 . The computer-implemented method of claim 2 , the computer-implemented method further comprising:
tuning, by the one or more processors, the machine-learning model based on the user feedback.
4 . The computer-implemented method of claim 1 , wherein a distribution profile defines the sensitivity score, and wherein the distribution profile includes a censorship profile for a specific country or a specific region.
5 . The computer-implemented method of claim 4 , the computer-implemented method further comprising:
storing, by the one or more processors, the distribution profile in the multidimensional tensor.
6 . The computer-implemented method of claim 5 , wherein the mapping the annotated plurality of frames to the sensitivity score includes:
utilizing, by the one or more processors, the multidimensional tensor to determine the sensitivity score for the sensitive event.
7 . The computer-implemented method of claim 1 , wherein the visual representation includes a heat map.
8 . A computer system for determining a sensitive portion of a dataset for media production, the computer system comprising:
a memory having processor-readable instructions stored therein; and
one or more processors configured to access the memory and execute the processor-readable instructions, which when executed by the one or more processors configures the one or more processors to perform a plurality of functions, including functions for:
receiving, by one or more processors, a digital dataset corresponding to a video source file;
processing, by the one or more processors, the digital dataset into a plurality of frames;
annotating, by the one or more processors via a machine-learning model, at least one of the plurality of frames with censorship metadata, wherein the censorship metadata includes a sensitive event;
mapping, by the one or more processors, utilizing a multidimensional tensor, the annotated plurality of frames to a sensitivity score based on the censorship metadata, wherein the mapping includes mapping the sensitive event to the sensitivity score;
assigning, by the one or more processors, a sensitivity tag to the sensitivity score, wherein the sensitivity tag indicates a censorship likelihood for the sensitive event;
updating, by the one or more processors, the multidimensional tensor to include the sensitivity tag, the multidimensional tensor having a plurality of dimensions including a sensitivity tag category, a localization region, and a sensitivity level; and
outputting, by the one or more processors, a visual representation of the sensitive event and the sensitivity tag to a display.
9 . The computer system of claim 8 , the functions further comprising:
receiving, by the one or more processors, user feedback indicating a sensitivity score accuracy for the sensitivity score.
10 . The computer system of claim 9 , the functions further comprising:
tuning, by the one or more processors, the machine-learning model based on the user feedback.
11 . The computer system of claim 8 , wherein a distribution profile defines the sensitivity score, and wherein the distribution profile includes a censorship profile for a specific country or a specific region.
12 . The computer system of claim 11 , the functions further comprising:
storing, by the one or more processors, the distribution profile in the multidimensional tensor.
13 . The computer system of claim 12 , wherein the mapping the annotated plurality of frames to the sensitivity score includes:
utilizing, by the one or more processors, the multidimensional tensor to determine the sensitivity score for the sensitive event.
14 . The computer system of claim 8 , wherein the visual representation includes a heat map.
15 . A non-transitory computer-readable medium containing instructions for determining a sensitive portion of a dataset for media production, the instructions comprising:
receiving a digital dataset corresponding to a video source file;
processing the digital dataset into a plurality of frames;
annotating, via a machine-learning model, at least one of the plurality of frames with censorship metadata, wherein the censorship metadata includes a sensitive event;
mapping, utilizing a multidimensional tensor, the annotated plurality of frames to a sensitivity score based on the censorship metadata, wherein the mapping includes mapping the sensitive event to the sensitivity score;
assigning a sensitivity tag to the sensitivity score, wherein the sensitivity tag indicates a censorship likelihood for the sensitive event;
updating the multidimensional tensor to include the sensitivity tag, the multidimensional tensor having a plurality of dimensions including a sensitivity tag category, a localization region, and a sensitivity level; and
outputting a visual representation of the sensitive event and the sensitivity tag to a display.
16 . The non-transitory computer-readable medium of claim 15 , the instructions further comprising:
receiving user feedback indicating a sensitivity score accuracy for the sensitivity score.
17 . The non-transitory computer-readable medium of claim 16 , the instructions further comprising:
tuning the machine-learning model based on the user feedback.
18 . The non-transitory computer-readable medium of claim 15 , wherein a distribution profile defines the sensitivity score, and wherein the distribution profile includes a censorship profile for a specific country or a specific region.
19 . The non-transitory computer-readable medium of claim 18 , the instructions further comprising:
storing and the distribution profile in the multidimensional tensor.
20 . The non-transitory computer-readable medium of claim 19 , wherein the mapping the annotated plurality of frames to the sensitivity score includes:
utilizing the multidimensional tensor to determine the sensitivity score for the sensitive event.