Automated Highlight Video Generation Using User Action Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating highlight videos from livestreams are inefficient, often requiring manual selection of important segments or automated processes that waste processing resources by not adequately accounting for user interaction data and segment relevance.
Innovation Solution
A computer-implemented system that automatically generates highlight videos by retrieving user action data, selecting relevant source videos, analyzing them to identify clip points, splitting them into segments, calculating segment highlight scores, and selecting segments for inclusion in a new video.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If highlight videos are created manually by going through livestream videos to identify important moments, then the quality and relevance of highlight segments can be ensured, but the complexity and time consumption of the process becomes too high to be efficient
Solution Approach 1:
The patent replaces the manual mechanical process of reviewing and selecting highlight segments with an automated computerized system. The system uses audio processing to detect voice activity and silence, video processing to analyze visual content, and user action data to identify engaging segments, thereby eliminating the need for manual video review while maintaining accurate segment selection.
Solution Approach 2:
The system enables the livestream processing system to automatically identify and select highlight segments without human intervention. By incorporating user action data (likes, shares, comments) and automated audio-video analysis, the system serves itself in determining which segments are most valuable, making the process efficient and scalable.
2Productivity
If highlight videos are created automatically by trimming predetermined periods and stitching together fixed-duration segments, then processing efficiency can be improved, but the relevance and quality of highlight segments deteriorates due to waste of processing resources on unpopular videos
Solution Approach 1:
The patent applies partial action by selectively processing only a subset of livestream videos - specifically those that meet certain criteria such as minimum duration, sufficient user engagement, or high viewer count. This prevents wasting processing resources on videos that are unlikely to produce valuable highlight content, while still maintaining high productivity for the selected videos through automated analysis.
Solution Approach 2:
The system dynamically adjusts processing parameters based on video characteristics and user engagement metrics. Videos with higher user action data receive more intensive analysis, while less popular videos may be processed with reduced complexity or skipped entirely, optimizing the balance between productivity and resource utilization.
3Reliability
If all livestream videos are processed to create highlight videos, then comprehensive coverage is achieved, but processing resources are wasted on videos with poor customer traction
Solution Approach 1:
The system performs preliminary filtering of livestream videos based on user action data and engagement metrics before initiating full highlight video processing. By pre-identifying videos that meet minimum thresholds for popularity and engagement, the system ensures reliable highlight video availability for worthwhile content while avoiding unnecessary processing of low-value videos.
4Measurement precision
If user action data is collected and analyzed to identify important segments, then the relevance of highlight videos improves, but the complexity of data collection and processing increases
Solution Approach 1:
The patent creates a multi-functional system that simultaneously performs multiple tasks: collecting user action data, processing audio to detect voice activity, analyzing video content, calculating highlight scores, and generating output videos. This universal system handles diverse functions through integrated modules, reducing overall system complexity compared to separate systems for each function.
Data Source
AI summary
Computer-implemented systems and methods for automatic generation of highlight videos are disclosed. The systems and methods may be configured to: retrieve user action data for a plurality of source videos from a predetermined time period; select, based on the retrieved user action data, a predetermined number of source videos of the plurality of source videos; retrieve the selected source videos; for each of the selected source videos: analyze the source video to identify clip points; split the analyzed source video into a plurality of segments; for each segment of the plurality of segments: determine segment data; and calculate a segment highlight score based on the determined segment data; select a number of segments from the plurality of segments based on the calculated segment highlight score of each segment; generate a new video based on the selected segments; and store the new video.