Jan Hendrik Metzen

This inventor holds 20 USPTO granted patents and 27 published patent applications and 11 EPO patents, primarily in Machine Learning Robustness. Top assignees: Robert Bosch, Carnegie Mellon University. Active years: 2019-2026.

IDiyas Innovation Intelligence. (2026). Inventor Profile: Jan Hendrik Metzen. Retrieved from https://idiyas.com/inventor/jan-hendrik-metzen

Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile

USPTO Granted Patents = 20 (Patent Applications = 27)

% Patents Active = 85.0

EPO Patents = 11 (Patent Applications = 27)

Average Co-Inventor Count = 2.4

ph-index = 2

Forward Citations = 49(Granted Patents)

Company Filing History:

Carnegie Mellon University

patents

where one patent can have more than one assignee

Years Active: 2019-2026

Loading Chart...

Areas of Expertise:

Adversarial Perturbations

Image Classification

Robustness Testing

Neural Networks

Semantic Segmentation

AI Vulnerability

Deep Learning

Classification Algorithms

Training Data Adaptation

Object Tracking

Automatic Learning Systems

Sensitivity Measurement

20 patents (USPTO):Sort by Forward CitationsExplore Patents

1. 12736933 - Device and method for determining adversarial perturbations of a machine learning system (2026-09-15),

2. 12705868 - Computer-implemented method and system for ascertaining a misclassification value during the classification of images into at least one image category by a classification algorithm (2026-08-11),

3. 12620200 - Method for operating a technical system and technical system (2026-05-05),

4. 12536339 - Device and method for determining adversarial patches for a machine learning system (2026-01-27),

5. 12456284 - Device and method for classifying images and accessing the robustness of the classification (2025-10-28),

6. 12340572 - Device and method to adapt a pretrained machine learning system to target data that has different distribution than the training data without the necessity of human annotations on target data (2025-06-24),

7. 12243283 - Device and method for determining a semantic segmentation and/or an instance segmentation of an image (2025-03-04),

8. 12205349 - System and method for improving robustness of pretrained systems in deep neural networks utilizing randomization and sample rejection (2025-01-21),

9. 12204653 - System and method, in particular computer implemented method for determining a perturbation for attacking and/or validating an object tracker (2025-01-21),

10. 12086214 - Method and device for creating a machine learning system (2024-09-10),

11. 12014280 - Measuring the sensitivity of neural network image classifiers against adversarial attacks (2024-06-18),

12. 11947625 - Data-based updating of the training of classifier networks (2024-04-02),

13. 11727277 - Method and apparatus for automatically producing an artificial neural network (2023-08-15),

14. 11676025 - Method, apparatus and computer program for generating robust automatic learning systems and testing trained automatic learning systems (2023-06-13),

15. 11599827 - Method and apparatus for improving the robustness of a machine learning system (2023-03-07),

Title: Jan Hendrik Metzen - A Visionary in the World of Inventions

Introduction:

Jan Hendrik Metzen, a brilliant inventor based in Boeblingen, DE, is known for his passion for innovation and dedication to pushing the boundaries of what is possible. With 11 patents to his name, he continues to inspire future generations of inventors to think boldly and create solutions that shape the world for the better.

Latest Patents:

1. Method and device for creating a machine learning system - This patent involves creating a machine learning system designed for segmentation and object detection in images. The method includes providing a directed graph, selecting a path through the graph, creating a machine learning system as a function of the selected path, and training the system.

2. Measuring the sensitivity of neural network image classifiers against adversarial attacks - This patent focuses on measuring the sensitivity of a classifier for digital images against adversarial attacks. The method includes providing a digital image, perturbing the image with a disturbance, determining the impact of the disturbance on the classification result, and optimizing parameters to maximize this impact.

Career Highlights:

Jan Hendrik Metzen has made significant contributions during his time at Robert Bosch G.m.b.H. and Carnegie Mellon University. His innovative work showcases his expertise in the field of invention and highlights his commitment to advancing technology for the benefit of society.

Collaborations:

Throughout his career, Jan Hendrik Metzen has collaborated with talented individuals such as Chaithanya Kumar Mummadi and Volker Fischer. These collaborations have led to groundbreaking inventions and have further solidified his reputation as a visionary inventor.

Conclusion:

In conclusion, Jan Hendrik Metzen's innovative spirit and dedication to invention make him a true visionary in the world of inventions. His impactful work not only showcases his talent but also serves as a source of inspiration for aspiring inventors worldwide. Jan Hendrik Metzen's legacy will continue to shape the future of technology and innovation for years to come.

Profile summary based on public USPTO records.