Video Summary Generation by Extracting and Overlaying Event Pixels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monitoring and reviewing video recordings is a time-consuming task, especially when searching for specific events without knowing the exact time, as increasing playback speed or using multiple windows can lead to missing important details or becoming overwhelmed with multiple streams.
Innovation Solution
Generating a summary video sequence by identifying event video sequences with objects of interest, extracting and overlaying their pixels while maintaining spatial and temporal relationships, and disregarding time periods without these objects, to create a condensed version that preserves concurrent actions and interactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If playback speed is increased to reduce review time, then time efficiency is improved, but important details may be missed
Solution Approach 1:
The patent extracts only the relevant portions of video containing objects of interest from the complete video sequence. By identifying and isolating event video sequences where objects of interest are present, the system creates a condensed summary that eliminates irrelevant time periods while preserving all important information, allowing reviewers to watch at normal speed without missing details.
Solution Approach 2:
The patent segments the complete video sequence into multiple event video sequences based on the presence and absence of objects of interest. Each segment contains only relevant content with objects of interest, and these segments are concatenated to form a compressed summary sequence, enabling efficient review without information loss.
2Productivity
If multiple windows are used to view different video portions simultaneously, then review speed is improved, but complexity of tracking events increases
Solution Approach 1:
The patent merges multiple event video sequences containing objects of interest into a single compressed summary sequence. By overlaying and concatenating relevant segments from different time periods into one continuous video stream, the system allows reviewers to view all important events in a single window, eliminating the need to track multiple simultaneous streams while maintaining review efficiency.
3Loss of time
If playback speed is increased by skipping frames, then review time is reduced, but reliability of event detection decreases
Solution Approach 1:
The patent extracts complete event video sequences containing objects of interest at their original playback speed from the compressed video sequence. By preserving the full frame sequences of relevant events rather than using accelerated or skipped frames, the system ensures high reliability in event detection while still achieving time savings by excluding irrelevant time periods from the summary.
Data Source
AI summary
A method for generating a summary video sequence from a source video sequence is disclosed. The method comprises: identifying, in the source video sequence, event video sequences, wherein each event video sequence comprises consecutive video frames in which one or more objects of interest are present; extracting, from video frames of one or more event video sequences of the event video sequences, pixels depicting the respective one or more objects of interest; while keeping spatial and temporal relations of the extracted pixels as in the source video sequence, overlaying the extracted pixels of the video frames of the one or more event video sequences onto video frames of a main event video sequence of the event video sequences, thereby generating the summary video sequence. A video processing device configured to generate the summary video sequence is also disclosed.