TY - CONF A1 - Bertovic, Marija T1 - Tutorial: NDE Reliability and the Influence of Human Factors N2 - Human factors significantly influence the reliability of non-destructive testing (NDT) systems, impacting safety, efficiency, and cost. This tutorial explores the interplay between human capabilities, organizational context, and technological systems in NDT inspections. It starts with a short overview of reliability models, including the Probability of Detection (POD) framework, alongside with advanced approaches for reliability assessment. The distinction between human factors and human error is emphasized, advocating a systemic approach to human error that addresses underlying organizational and environmental conditions. The implications of increasing automation and AI integration are discussed, highlighting the challenges of trust, usability, and skill preservation in human-AI collaboration. Practical strategies, including user-centered design, targeted training, and risk management frameworks, are proposed to enhance inspection performance and maintain NDT reliability in evolving technological landscapes. This tutorial underscores the critical role of human factors in the transition to NDT 4.0 and the development of effective human-machine systems. T2 - USES2 Training week#2 CY - Berlin, Germany DA - 24.06.2024 KW - Human Factors KW - Non-Destructive Testing KW - Reliability KW - Human-Machine Interaction KW - Artificial Intelligence PY - 2024 AN - OPUS4-61874 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Huo, Wenjie A1 - Bakir, Nasim A1 - Gumenyuk, Andrey A1 - Rethmeier, Michael A1 - Wolter, Katinka T1 - Detection of solidification crack formation in laser beam welding videos of sheet metal using neural networks N2 - AbstractLaser beam welding has become widely applied in many industrial fields in recent years. Solidification cracks remain one of the most common welding faults that can prevent a safe welded joint. In civil engineering, convolutional neural networks (CNNs) have been successfully used to detect cracks in roads and buildings by analysing images of the constructed objects. These cracks are found in static objects, whereas the generation of a welding crack is a dynamic process. Detecting the formation of cracks as early as possible is greatly important to ensure high welding quality. In this study, two end-to-end models based on long short-term memory and three-dimensional convolutional networks (3D-CNN) are proposed for automatic crack formation detection. To achieve maximum accuracy with minimal computational complexity, we progressively modify the model to find the optimal structure. The controlled tensile weldability test is conducted to generate long videos used for training and testing. The performance of the proposed models is compared with the classical neural network ResNet-18, which has been proven to be a good transfer learning model for crack detection. The results show that our models can detect the start time of crack formation earlier, while ResNet-18 only detects cracks during the propagation stage. KW - Artificial Intelligence KW - Software PY - 2023 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-586116 DO - https://doi.org/10.1007/s00521-023-09004-y SN - 0941-0643 VL - 35 IS - 34 SP - 24315 EP - 24332 PB - Springer Science and Business Media LLC AN - OPUS4-58611 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Rühle, Bastian A1 - Hodoroaba, Vasile-Dan T1 - Automatic Image Segmentation and Analysis using Neural Networks N2 - We present a workflow for obtaining fully trained artificial neural networks that can perform automatic particle segmentations of agglomerated, non-spherical nanoparticles from electron microscopy images “from scratch”, without the need for large training data sets of manually annotated images. This is achieved by using unsupervised learning for most of the training dataset generation, making heavy use of generative adversarial networks and especially unpaired image-to-image translation via cycle-consistent adversarial networks. The whole process only requires about 15 minutes of hands-on time by a user and can typically be finished within less than 12 hours when training on a single graphics card (GPU). After training, SEM image analysis can be carried out by the artificial neural network within seconds, and the segmented images can be used for automatically extracting and calculating various other particle size and shape descriptors. T2 - Machine Learning Workshop CY - Online Meeting DA - 18.03.2021 KW - Electron Microscopy KW - Neural Networks KW - Artificial Intelligence KW - Image Segmentation KW - Automated Image Analysis PY - 2021 AN - OPUS4-52304 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -