Video Searching System Using Graphical Event Timeline
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video surveillance systems are inefficient in analyzing and searching through long video recordings, as they only allow for quick mode playback and cannot effectively display key contents or provide temporal relationships between objects of interest.
Innovation Solution
A video searching method and system that sets a region-of-interest and capturing conditions to capture object snapshots of moving objects, integrating them into a graphical event timeline, allowing for the display of object behaviors over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional video displaying system displays video in quick mode at time axis, then video playback speed is improved, but user cannot watch all objects-of-interest at different time points and cannot understand temporal relationships
Solution Approach 1:
The patent segments the video content by extracting and displaying multiple snapshots at different time points simultaneously on the screen. Each snapshot represents a specific moment in the video, allowing users to view multiple temporal moments at once rather than sequentially, thus preserving temporal relationship information while maintaining fast access to different time points.
Solution Approach 2:
The patent adds a spatial dimension to time representation by arranging snapshots from different time points in a spatial layout on the screen. This transforms the one-dimensional time axis navigation into a two-dimensional spatial arrangement, enabling users to perceive temporal relationships through spatial positioning of snapshots.
2Loss of information
If conventional video displaying system displays complete video screens according to time sequence, then all video contents are displayed, but screen display time segment is very short and non-key portion cannot be excluded, limiting searching efficiency
Solution Approach 1:
The patent extracts key information from the video by capturing and displaying representative snapshots at critical time points. Instead of showing complete video screens including non-key portions, the system extracts essential moments and presents them in a condensed format, maintaining video content completeness while eliminating redundant information to improve searching efficiency.
Solution Approach 2:
The patent applies partial action by displaying only the most relevant snapshots rather than the entire video content. By selecting and presenting key frames that represent important events or changes, the system provides sufficient information for effective video searching without requiring users to review all video content.
3Ease of operation
If user watches video screen at specific time point in conventional system, then instant video information is obtained, but user cannot understand moving behaviors of objects across multiple time points
Solution Approach 1:
The patent merges multiple snapshots from different time points into a single unified display interface. By combining these temporal instances in one view, users can simultaneously observe object positions and behaviors across multiple time points, understanding movement patterns without switching between different time point views, thus maintaining operational simplicity while preserving movement information.
Data Source
AI summary
A video searching method and a video searching system are applied to an auxiliary searching of an original video. The video searching method includes setting a region-of-interest in a screen of the original video, setting a capturing condition that includes an interactive relation between the region-of-interest and a moving object of the original video, capturing a background screen of the original video and object snapshots of the moving object that comply with the capturing condition, selecting at least one of the object snapshots captured from the same moving object as a representative object snapshot, or integrating some of the object snapshots as a single representative object snapshot; and pasting the representative object snapshot at a coordinate corresponding to the background screen according to a corresponding timestamp, to generate a graphical event timeline that displays all the object snapshots.