Automated Defect Detection for Industrial Quality Control

 

Motivation

The use of machine learning/deep learning in industrial image-based quality control holds enormous potential, but is currently hindered by two circumstances:

  • Supervised learning methods require large amounts of data, also from error-prone states, which have to be collected at high cost..
  • No procedures exist that are suitable for managing the product life cycle of developed processes.

The methods and current developments in the field of Open Set Recognition offer possible solutions and are being researched in the project..

Scientific Questions

To overcome the problem of large data sets, anomaly detection algorithms are being researched and developed. In contrast to classical supervised learning methods, the goal of anomaly detection is a descriptive and discriminative description of the normal state.
Algorithms are developed on the public MVTec dataset and also adapted to special applications, such as fault detection in complex textiles (AiF project OnLoomPattern).

As a basis for the development of methods for product lifecycle management, methods of out-of-distribution (OOD) detection are being researched and further developed. In combination with explainble AI (XAI) algorithms, these allow on the one hand the automatic assessment of whether a machine learning system may be used (OOD), and on the other hand the verification of the automatic assessment (XAI). Furthermore, the algorithms are adapted to the special use case of the forgery-resistant animal hair fiber determination (AiF project KiT).

Theses

New theses are regularly advertised in the area of Automated Defect Detection for Industrial Quality Control. In addition to the general overview, there are also numerous topics that have not yet been advertised, which will be gladly presented in a personal conversation.

Partners

External Funding

Contact

Publications

2022

Rippel, Oliver, Zwinge, Corinna and Merhof, Dorit
Increasing the Generalization of Supervised Fabric Anomaly Detection Methods to Unseen Fabrics
In: Sensors 22 (13)


2022

Oliver Rippel, Nikolaj Schönfelder, Khosrow Rahimi, Juliana Kurniadi, Andreas Herrmann and Dorit Merhof
Panoptic Segmentation of Animal Fibers
In: 2022 IEEE International Instrumentation and Measurement Technology Conference (I2MTC)


2022

Oliver Rippel, Arnav Chavan, Chucai Lei and Dorit Merhof
Transfer Learning Gaussian Anomaly Detection by Fine-tuning Representations
In: Proceedings of the 2nd International Conference on Image Processing and Vision Engineering - IMPROVE


2022

Oliver Rippel, Sergen Gülçelik, Khosrow Rahimi, Juliana Kurniadi, Andreas Herrmann and Dorit Merhof
Animal Fiber Identification under the Open Set Condition
In: 17th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISAPP)


2021

Oliver Rippel, Patrick Mertens, Eike König and Dorit Merhof
Gaussian Anomaly Detection by Modeling the Distribution of Normal Data in Pre-Trained Deep Features
In: IEEE Transactions on Instrumentation and Measurement


2021

Oliver Rippel and Dorit Merhof
Leveraging pre-trained Segmentation Networks for Anomaly Segmentation
In: IEEE 26th International Conference on Emerging Technologies and Factory Automation (ETFA)


2021

Oliver Rippel, Peter Haumering, Johannes Brauers and Dorit Merhof
Anomaly Detection for the Automated Visual Inspection of PET Preform Closures
In: IEEE 26th International Conference on Emerging Technologies and Factory Automation (ETFA)


2021

Oliver Rippel, Niclas Bilitewski, Khosrow Rahimi, Juliana Kurniadi, Andreas Herrmann and Dorit Merhof
Identifying Pristine and Processed Animal Fibers using Machine Learning
In: IEEE International Instrumentation and Measurement Technology Conference (I2MTC)


2021

Oliver Rippel, Maximilian Müller, Andreas Münkel, Thomas Gries and Dorit Merhof
Estimating the Probability Density Function of new Fabrics for Fabric Anomaly Detection
In: 10th International Conference on Pattern Recognition Applications and Methods (ICPRAM)


2021

Oliver Rippel, Patrick Mertens and Dorit Merhof
Modeling the Distribution of Normal Data in Pre-Trained Deep Features for Anomaly Detection
In: 25th International Conference on Pattern Recognition (ICPR)


2020

Oliver Rippel, Maximilian Müller and Dorit Merhof
GAN-based Defect Synthesis for Anomaly Detection in Fabrics
In: IEEE 25th International Conference on Emerging Technologies and Factory Automation (ETFA)