MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641096282 A) filed by Saveetha Engineering College on August 10, 2026, for Self-Learning Machine Learning System For Industrial Quality Inspection.
Inventors include Mr. P. Sankar; Dr. Dr. J. Anish Kumar; Ms. Lakshanya N; Ms. Abisha Linu L; and Ms. Parveen Sulthana J.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: The present invention relates to a self-learning machine learning system for automated industrial quality inspection of manufactured products, components, assemblies, and materials. The system comprises an inspection acquisition subsystem configured to acquire inspection data from an article using one or more sensing devices, a processing subsystem configured to preprocess and extract relevant inspection characteristics, and a machine learning subsystem configured to classify the article according to a quality condition. The system determines a confidence value associated with each inspection result and identifies uncertain, novel, or potentially misclassified inspection conditions for further verification. Verified inspection information obtained from an operator, secondary inspection apparatus, functional testing arrangement, or other trusted source is associated with corresponding inspection data and stored in a learning repository. A self-learning subsystem selectively identifies informative samples based on model uncertainty, classification errors, defect novelty, recurring misclassification, or changes in inspection-data distribution and uses the selected samples to adapt a machine learning model.
Disclaimer: Curated by HT Syndication.
