EPO Patent Application: Method for Correcting Ground Truth Labels in Training Datasets
Summary
The European Patent Office has published patent application EP4711990A1 detailing a method for correcting ground truth labels in training datasets for machine learning models used in computer vision tasks. The application, filed by AUMOVIO Germany GmbH, describes a process involving training a model, modifying sample images, obtaining predicted labels, and then correcting the ground truth labels based on a comparison.
What changed
European Patent Office (EPO) patent application EP4711990A1, published on March 18, 2026, outlines a novel method for refining the accuracy of ground truth labels used in training machine learning models for computer vision. The invention, assigned to AUMOVIO Germany GmbH, involves an iterative process where a model is trained, sample images are modified, and the model's predictions are compared against existing ground truth labels to identify and correct inaccuracies. This approach aims to improve the reliability and performance of AI models by enhancing the quality of their training data.
This patent application, while not imposing direct regulatory obligations on companies, signifies a technological advancement in AI development. Companies involved in developing or utilizing AI for computer vision tasks, particularly those focused on data quality and model training, should be aware of this patented methodology. It may influence future licensing agreements or development strategies for AI training data correction tools. No immediate compliance actions are required, but it highlights an area of innovation that could become relevant for intellectual property considerations in the AI sector.
Source document (simplified)
A METHOD, APPARATUS, AND COMPUTER-READABLE MEDIUM FOR CORRECTING GROUND TRUTH LABELS IN TRAINING DATASETS
Publication EP4711990A1 Kind: A1 Mar 18, 2026
Applicants
AUMOVIO Germany GmbH
Inventors
Kassem Sbeyti, Moussa, Nowzad, Azarm, Wirth, Christian
Abstract
Embodiments of the present disclosure relate to a method for correcting ground truth labels used in training a machine learning model for a computer vision task. A training method and ground truth labels used for training a machine learning model are also disclosed. A machine learning model, a computer program, a computer-readable data carrier and an apparatus are also presented. The method for correcting ground truth labels comprises obtaining sample images and ground truth labels for the sample images, training the machine learning model using the sample images and the ground truth labels, modifying the sample images, obtaining, by the machine learning model, predicted labels for the modified sample images, and correcting the ground truth labels based on a comparison of the ground truth labels with the predicted labels.
IPC Classifications
G06N 20/00 20190101AFI20250522BHEP
Designated States
AL, AT, BE, BG, CH, CY, CZ, DE, DK, EE, ES, FI, FR, GB, GR, HR, HU, IE, IS, IT, LI, LT, LU, LV, MC, ME, MK, MT, NL, NO, PL, PT, RO, RS, SE, SI, SK, SM, TR
Named provisions
Related changes
Source
Classification
Who this affects
Taxonomy
Browse Categories
Get Telecom & Technology alerts
Weekly digest. AI-summarized, no noise.
Free. Unsubscribe anytime.
Get alerts for this source
We'll email you when ChangeBridge: EPO Bulletin - AI & Computing (G06N) publishes new changes.