Multi-Camera Re-Identification Calibration Using Pairwise Thresholds

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing multi-camera systems lack an effective method for accurately determining the re-identification threshold, leading to high false positives or false negatives in identifying individuals across multiple cameras.

Innovation Solution

A method for calibrating an object re-identification solution that involves detecting objects in pairs of cameras, computing normalized distances between digital signatures, and determining a re-identification threshold for each camera pair, taking into account differences in field of view and lighting, to reduce false positives and negatives.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a single re-identification threshold is used for all cameras in the array, then the system operation is simplified, but the measurement precision of re-identification decreases due to differences in field of view and lighting between cameras

Engineering Contradiction:

Improvesystem operation simplicityVSAvoidre-identification threshold accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The patent divides the re-identification threshold determination into separate pairs of cameras. Instead of using a single global threshold for all cameras, the system computes individual thresholds for each camera pair based on their specific characteristics (field of view, lighting conditions). This segmentation allows each threshold to be optimized for its specific camera pair, resolving the contradiction between operational simplicity and measurement precision.

Inventive Principle:

Principle #1Segmentation

2Reliability

If the re-identification threshold is set high to reduce false negatives, then more individuals are correctly identified, but false positives increase

Engineering Contradiction:

Improvere-identification accuracyVSAvoidfalse positives

Core Design Contradiction:

ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent applies local quality by computing distinct re-identification thresholds for each camera pair rather than using a uniform threshold across all cameras. Each threshold is locally optimized based on the specific characteristics of that camera pair (field of view, lighting conditions), allowing the system to minimize false positives for each pair while maintaining high re-identification accuracy. This local optimization resolves the contradiction between reducing false negatives and avoiding false positives.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS12518504B2Method for calibrating an object re-identification solution implementing an array of a plurality of cameras

Publication Date: 2026.01.06 BULL SA

AI summary

The invention includes a method for calibrating an object re-identification solution implementing an array of several cameras. For at least one pair of two cameras from among the several camera, the method includes detecting objects in images taken by each of the two cameras, and computing a normalized distance between a digital signature of each object detected by one of the two cameras, and that of each different object detected by the other one of the two cameras. The method also includes determining, as a function of said normalized distances, a distance threshold, called re-identification threshold, that will be used for re-identifying objects in the images taken by the two cameras. The invention also includes a computer program and a device implementing such a calibration method, and a method and a system for re-identifying individuals calibrated by such a calibration method.