Sistema de Reconocimiento Facial Federado en Conjuntos Abiertos basado en OpenMax

2025-08-26

AI Summary

To address privacy preservation and unknown-identity discrimination in open-set face recognition under federated learning, this paper proposes a federated open-set recognition framework integrating OpenMax. The method computes class-wise mean activation vectors and intra-class distances locally at clients, while the server aggregates activation features to construct a scalable open-set discrimination mechanism—eliminating raw-data transmission and thus ensuring privacy without compromising unknown-class detection capability. Crucially, OpenMax is innovatively adapted to the federated paradigm to enable distributed estimation of unknown-identity confidence scores. Extensive experiments on multiple benchmark datasets demonstrate that the proposed approach achieves significant improvements in known/unknown face discrimination accuracy (e.g., +8.2% AUROC) while maintaining strong privacy guarantees. Moreover, it exhibits superior robustness compared to existing federated closed-set and centralized open-set methods.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Bias, Fairness & Privacy
• Search and Optimization: Local Search

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Facial recognition powered by Artificial Intelligence has achieved high accuracy in specific scenarios and applications. Nevertheless, it faces significant challenges regarding privacy and identity management, particularly when unknown individuals appear in the operational context. This paper presents the design, implementation, and evaluation of a facial recognition system within a federated learning framework tailored to open-set scenarios. The proposed approach integrates the OpenMax algorithm into federated learning, leveraging the exchange of mean activation vectors and local distance measures to reliably distinguish between known and unknown subjects. Experimental results validate the effectiveness of the proposed solution, demonstrating its potential for enhancing privacy-aware and robust facial recognition in distributed environments. -- El reconocimiento facial impulsado por Inteligencia Artificial ha demostrado una alta precisión en algunos escenarios y aplicaciones. Sin embargo, presenta desafíos relacionados con la privacidad y la identificación de personas, especialmente considerando que pueden aparecer sujetos desconocidos para el sistema que lo implementa. En este trabajo, se propone el diseño, implementación y evaluación de un sistema de reconocimiento facial en un escenario de aprendizaje federado, orientado a conjuntos abiertos. Concretamente, se diseña una solución basada en el algoritmo OpenMax para escenarios de aprendizaje federado. La propuesta emplea el intercambio de los vectores de activación promedio y distancias locales para identificar de manera eficaz tanto personas conocidas como desconocidas. Los experimentos realizados demuestran la implementación efectiva de la solución propuesta.

Problem

• Federated learning for facial recognition in open-set scenarios
• Distinguishing known and unknown subjects using OpenMax algorithm
• Enhancing privacy-aware recognition in distributed environments

Innovation

• Federated learning framework for open-set scenarios
• Integrates OpenMax algorithm for unknown subject identification
• Exchanges mean activation vectors and local distances