Duplicate Video Detection via Multi-Feature Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video comparison and de-duplication methods are inefficient, with low speed, poor accuracy, and high resource consumption.
Innovation Solution
A method and device for determining duplicate videos by extracting multiple types of video features, performing sequence comparisons, and ranking videos based on similarity scores to identify duplicate video segments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing video comparison methods are used, then duplicate video detection can be performed, but the processing speed is low and accuracy is poor
Solution Approach 1:
The patent segments video comparison into multiple feature dimensions (color, texture, motion, audio, etc.) and processes each dimension separately through dedicated comparison modules. This segmentation allows parallel processing of different feature types, improving both speed and accuracy by specializing each module for its specific feature type rather than using a single general-purpose comparison method
Solution Approach 2:
The patent transitions from traditional single-dimensional video comparison to multi-dimensional feature space comparison by extracting and comparing videos across multiple feature dimensions simultaneously. This dimensional expansion enables more comprehensive video similarity assessment, resolving the accuracy-speed contradiction by processing multiple dimensions in parallel
2Measurement precision
If comprehensive video comparison is performed, then detection accuracy improves, but resource consumption increases
Solution Approach 1:
By dividing the comprehensive video comparison task into segmented feature extraction and comparison modules (color feature module, texture feature module, motion feature module, etc.), the patent enables selective and parallel processing of only relevant features, reducing overall resource consumption while maintaining comprehensive comparison accuracy
Solution Approach 2:
The patent implements a hierarchical comparison approach where not all feature dimensions are processed with equal depth for every video pair. The system performs partial comprehensive comparison by adjusting the level of detail in each feature dimension based on initial screening results, reducing unnecessary resource expenditure on already-identified duplicates
3Reliability
If traditional video de-duplication methods are used, then duplicate videos can be identified, but storage resource consumption is large
Solution Approach 1:
The patent extracts essential video features (color histograms, texture descriptors, motion vectors, audio fingerprints) and stores only these extracted feature representations rather than storing complete video files for comparison. This extraction approach maintains high identification reliability through comprehensive feature analysis while dramatically reducing storage requirements by storing compact feature vectors instead of full video content
Data Source
AI summary
The present invention relates to a method and a device for determining a duplicate video, wherein the method comprises: acquiring multiple types of video features of a query video; according to the multiple types of video features of the query video, sequentially comparing a plurality of existing videos respectively to obtain a sequence comparison result; performing a first ranking on the plurality of existing videos according to the sequence comparison result, and taking first n existing videos as first candidate videos according to a result of the first ranking, where n is a positive integer; and determining a duplication level of the query video according to the sequence comparison result of the first candidate videos.