Lifelong Deep Neural Networks for Industrial Quality Control Anomaly Detection

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 31st, 2026

Detected March 31st, 2026

Summary

USPTO granted Patent US12591227B2 to Neurala, Inc. covering Lifelong Deep Neural Network (L-DNN) technology for anomaly recognition in industrial quality control. The patent discloses a hybrid semi-supervised neural architecture combining DNN training precision on known classes with sensitivity to unknown classes, enabling defect detection with limited unbalanced training data. The patent contains 32 claims.

View original document View source feed page

What changed

USPTO issued Patent US12591227B2 to Neurala, Inc. on March 31, 2026, covering systems and methods for anomaly recognition using Lifelong Deep Neural Networks (L-DNN). The technology addresses industrial quality control challenges where good product data is abundant but bad product data is scarce. L-DNN combines DNN training on known classes with sensitivity to unknown class variations, enabling real-time learning from rare defect encounters post-deployment. The patent names eight inventors including Massimiliano Versace and was filed July 11, 2022 under Application No. 17811779.

Patent holders should review the granted claims to understand scope of exclusivity. Manufacturers implementing AI-based quality control systems should conduct freedom-to-operate analysis to ensure their systems do not infringe the granted claims. Legal teams should update IP portfolios to account for this issued patent in competitive landscapes for industrial AI inspection technologies.

What to do next

Review Patent US12591227B2 claims to assess scope of L-DNN exclusivity
Conduct freedom-to-operate analysis for AI quality control systems
Update competitive IP landscape for industrial AI anomaly detection

Source document (simplified)

← USPTO Patent Grants

Systems and methods for anomaly recognition and detection using lifelong deep neural networks

Grant US12591227B2 Kind: B2 Mar 31, 2026

Assignee

Neurala, Inc.

Inventors

Carl Palme, Carly Franca, Graham Voysey, Massimiliano Versace, Santiago Olivera, Vesa Tormanen, Alireza Majidi, Yiannis Papadopoulos

Abstract

Industrial quality control is challenging for artificial neural networks (ANNs) and deep neural networks (DNNs) because of the nature of the processed data: there is an abundance of consistent data representing good products, but little data representing bad products. In quality control, the task is changed from conventional DNN task of “recognize what I learned best” to “recognize what I have never seen before.” Lifelong DNN (L-DNN) technology is a hybrid semi-supervised neural architecture that combines the ability of DNNs to be trained, with high precision, on known classes, while being sensitive to any number of unknown classes or class variations. When used for industrial inspection, L-DNN exploits its ability to learn with little and highly unbalanced data. L-DNN's real-time learning capability takes advantage of rare cases of poor-quality products that L-DNN encounters after deployment. L-DNN can be applied to industrial inspections and manufacturing quality control.