Feature Vector Clustering for Surveillance Search Speed
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Camera surveillance systems face challenges in efficiently reviewing and identifying individuals or objects across multiple video feeds due to the large number of cameras generating separate video feeds, making the process cumbersome, time-consuming, and expensive.
Innovation Solution
A method involving the generation of feature vectors from images, clustering, filtering, and comparing these vectors to identify similar images within a database, allowing for efficient retrieval and display of relevant video footage based on user input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual review of multiple video feeds is performed, then identification of individuals or objects of interest can be achieved, but the process becomes cumbersome, time-consuming and expensive
Solution Approach 1:
The patent replaces manual mechanical review of video feeds with an automated computer-based system that uses image processing and feature vector comparison algorithms to automatically identify and locate individuals or objects of interest across multiple video feeds, thereby eliminating the time-consuming manual review process while maintaining identification accuracy
Solution Approach 2:
The patent creates feature vector representations (digital copies) of individuals or objects from video feeds, which can then be quickly compared and matched against other video feeds without requiring actual manual viewing of the entire video content, significantly reducing review time while preserving identification capability
2Reliability
If feature vectors are compared across all images in the database, then comprehensive search results are obtained, but the search process becomes computationally expensive and slow
Solution Approach 1:
The patent segments the large database of images into smaller manageable groups or batches, processing and comparing feature vectors in segmented portions rather than all at once, which maintains comprehensive search coverage while improving computational efficiency and reducing processing time
Solution Approach 2:
The patent performs preliminary processing of video feeds to extract and store feature vectors in advance, so that when a search is needed, the system can quickly compare against pre-processed data rather than processing raw video content in real-time, thereby maintaining search completeness while significantly improving search speed
Data Source
AI summary
A method and system for processing images for a search is provided, including: receiving a plurality of images selected from search results; for each image in the plurality of images, retrieving a feature vector associated with the image; selecting a subset of the feature vectors based on similarity of feature vectors associated with the images in the plurality of images; and performing a search for feature vectors in a database similar to the feature vectors in the subset of feature vectors.