Clustering-Based Person Re-Identification for Dynamic Identity Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional person re-identification techniques face challenges in dynamically updating and managing a database of recognized identities in large-scale distributed multi-camera systems, particularly when handling live video feeds from multiple cameras where individuals appear in multiple frames and multiple persons are detected simultaneously.
Innovation Solution
The implementation of a clustering process, such as incremental clustering, to assign identities to detected persons, allowing for the generation and updating of identity clusters, incorporating relational metrics and co-occurrence probabilities to enhance person re-identification, and dynamically manage the identity database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional person re-identification techniques are used, then basic identity recognition can be achieved, but the system fails to dynamically update and manage identity databases in large-scale distributed multi-camera systems
Solution Approach 1:
The patent implements dynamic database updating through incremental clustering that continuously processes new captured images and updates identity assignments in real-time. The system dynamically adjusts identity clusters based on new data from multiple cameras, enabling the database to adapt to changing scenarios while maintaining recognition accuracy through relational metrics and co-occurrence probabilities.
2Area of stationary object
If multiple camera feeds are processed simultaneously, then coverage of large-scale distributed systems is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent segments the identity database into multiple clusters, each representing a group of images associated with a specific identity. This segmentation allows the system to process multiple camera feeds independently and manage computational complexity by organizing data into manageable clusters rather than processing all images uniformly across the entire database.
Solution Approach 2:
The system pre-calculates and stores relational metrics and co-occurrence probabilities during the clustering process, enabling faster identity assignment when new images are captured. This preliminary computation reduces real-time processing requirements and allows the system to handle multiple camera feeds efficiently.
3Adaptability or versatility
If identity databases are continuously updated, then system adaptability is improved, but errors and noise in similarity scores increase
Solution Approach 1:
The patent incorporates feedback mechanisms where captured images are continuously evaluated against existing identity clusters using relational metrics. The system uses co-occurrence probabilities to verify identity assignments and adjusts clustering decisions based on feedback from new data, thereby maintaining measurement precision while enabling continuous adaptation.
Data Source
AI summary
Presented herein are techniques for assignment of an identity to a group of captured images. A plurality of captured images that each include an image of at least one person are obtained. For each of the plurality of captured images, relational metrics indicating a relationship between the image of the person in a respective captured image and the images of the persons in each of the remaining plurality of captured images is calculated. Based on the relational metrics, a clustering process is performed to generate one or more clusters from the plurality of captured images. Each of the one or more clusters are associated with an identity of an identity database. The one or more clusters may each be associated with an existing identity of the identity database or an additional identity that is not yet present in the identity database.