System, method and computer-readable medium for categorizing live streaming data
The present disclosure relates to a system, a method and a computer-readable medium for tagging live streaming data. The method includes generating a first intermediate tag for the live streaming program, generating a second intermediate tag for the live streaming program, and determining a final tag for the live streaming program according to the first intermediate tag and the second intermediate tag. The present disclosure can categorize contents in a more granular and precise way.
1. A method for tagging a live streaming program, comprising:
generating a first intermediate tag for the live streaming program;
generating a second intermediate tag for the live streaming program;
determining a final tag for the live streaming program according to the first intermediate tag and the second intermediate tag;
determining the first intermediate tag to have been generated consecutively for a first predetermined number of times;
determining the second intermediate tag to have been generated consecutively for a second predetermined number of times,
wherein the consecutive generation of the first intermediate tag for the first predetermined number of times and the consecutive generation of the second intermediate tag for the second predetermined number of times corresponds to the final tag; and
determining the generation of the first intermediate tag and the generation of the second intermediate tag to have met a predetermined criterion for determination of the final tag,
wherein the predetermined criterion requires a predetermined arrangement of the first intermediate tag and the second intermediate tag within a predetermined period of time,
wherein generating the first intermediate tag comprises an object recognizing process, a music recognizing process, an activity recognizing process, a motion recognizing process, a background recognizing process or a topic recognizing process.
2. The method according to claim 1 , wherein the generating the first intermediate tag includes inputting data of the live streaming program to a first machine learning model to obtain the first intermediate tag as an output from the first machine learning model,
wherein the generating the second intermediate tag includes inputting data of the live streaming program to a second machine learning model to obtain the second intermediate tag as an output from the second machine learning model,
wherein the generation of the first intermediate tag and the second intermediate tag is performed concurrently with provision of the live streaming program to users.
3. The method according to claim 1 , further comprising:
defining a first contribution score for the first intermediate tag with respect to the final tag;
defining a second contribution score for the second intermediate tag with respect to the final tag,
wherein determining the final tag comprises calculating a weighted sum of the first contribution score and the second contribution score.
4. The method according to claim 3 , wherein the first contribution score increases as the first intermediate tag is generated more frequently.
5. The method according to claim 1 , further comprising:
generating a plurality of intermediate tags for the live streaming program;
defining a contribution score for each intermediate tag with respect to each one of a plurality of candidate tags;
calculating a final score for each candidate tag according to the contribution score of each intermediate tag with respect to each candidate tag; and
determining the final tag to be the candidate tag with the highest final score.
6. A system for tagging a live streaming program, comprising one or a plurality of processors, wherein the one or plurality of processors execute a machine-readable instruction to perform:
generating, by the one or plurality of processors, a first intermediate tag for the live streaming program;
generating, by the one or plurality of processors, a second intermediate tag for the live streaming program; and
determining, by the one or plurality of processors, a final tag for the live streaming program according to the first intermediate tag and the second intermediate tag:,
determining, by the one or plurality of processors, the first intermediate tag to have been generated consecutively for a first predetermined number of times;
determining, by the one or plurality of processors, the second intermediate tag to have been generated consecutively for a second predetermined number of times,
wherein the consecutive generation of the first intermediate tag for the first predetermined number of times and the consecutive generation of the second intermediate tag for the second predetermined number of times corresponds to the final tag; and
determining, by the one or plurality of processors, the generation of the first intermediate tag and the generation of the second intermediate tag to have met a predetermined criterion for determination of the final tag,
wherein the predetermined criterion requires a predetermined arrangement of the first intermediate tag and the second intermediate tag within a predetermined period of time,
wherein generating the first intermediate tag comprises an object recognizing process, a music recognizing process, an activity recognizing process, a motion recognizing process, a background recognizing process or a topic recognizing process.
7. A non-transitory computer-readable medium including a program for tagging a live streaming program, wherein the program causes one or a plurality of computers to execute:
generating a first intermediate tag for the live streaming program;
generating a second intermediate tag for the live streaming program; and
determining a final tag for the live streaming program according to the first intermediate tag and the second intermediate tag;
determining the first intermediate tag to have been generated consecutively for a first predetermined number of times;
determining the second intermediate tag to have been generated consecutively for a second predetermined number of times,
wherein the consecutive generation of the first intermediate tag for the first predetermined number of times and the consecutive generation of the second intermediate tag for the second predetermined number of times corresponds to the final tag; and
determining the generation of the first intermediate tag and the generation of the second intermediate tag to have met a predetermined criterion for determination of the final tag,
wherein the predetermined criterion requires a predetermined arrangement of the first intermediate tag and the second intermediate tag within a predetermined period of time,
wherein generating the first intermediate tag comprises an object recognizing process, a music recognizing process, an activity recognizing process, a motion recognizing process, a background recognizing process or a topic recognizing process.