1 - 10 of 284
-
-
-
PublicationPräventionsmaßnahmen der digitalen Arbeit(Fraunhofer FIT, 2021)
-
PublicationElectricity Market Design 2030-2050: Moving Towards Implementation(Fraunhofer FIT, 2021)
-
PublicationDeepKneeExplainer: Explainable knee osteoarthritis diagnosis from radiographs and magnetic resonance imaging( 2021)Osteoarthritis (OA) is a degenerative joint disease, which significantly affects middle-aged and elderly people. Although primarily identified via hyaline cartilage change based on medical images, technical bottlenecks like noise, artifacts, and modality impose an enormous challenge on high-precision, objective, and efficient early quantification of OA. Owing to recent advancements, approaches based on neural networks (DNNs) have shown outstanding success in this application domain. However, due to nested non-linear and complex structures, DNNs are mostly opaque and perceived as black-box methods, which raises numerous legal and ethical concerns. Moreover, these approaches do not have the ability to provide the reasoning behind diagnosis decisions in the way humans would do, which poses an additional risk in the clinical setting. In this paper, we propose a novel explainable method for knee OA diagnosis based on radiographs and magnetic resonance imaging (MRI), which we called DeepKneeExplainer. First, we comprehensively preprocess MRIs and radiographs through the deep-stacked transformation technique against possible noises and artifacts that could contain unseen images for domain generalization. Then, we extract the region of interests (ROIs) by employing U-Net architecture with ResNet backbone. To classify the cohorts, we train DenseNet and VGG architectures on the extracted ROIs. Finally, we highlight class-discriminating regions using gradient-guided class activation maps (Grad-CAM++) and layer-wise relevance propagation (LRP), followed by providing human-interpretable explanations of the predictions. Comprehensive experiments based on the multicenter osteoarthritis study (MOST) cohorts, our approach yields up to 91% classification accuracy, outperforming comparable state-of-the-art approaches. We hope that our results will encourage medical researchers and developers to adopt explainable methods and DNN-based analytic pipelines towards an increasing acceptance and adoption of AI-assisted applications in the clinical practice for improved knee OA diagnoses.
-
PublicationGAIA-X and IDS(International Data Spaces Association, 2021)
-
PublicationFraunhofer-Institut für Angewandte Informationstechnik. Jahresbericht 2020(Fraunhofer FIT, 2021)
-
PublicationTowards Distributed Healthcare Systems - Virtual Data Pooling Between Cancer Registries as Backbone of Care and Research( 2021)German cancer registries offer a systematic approach for the collection, storage, and management of data on patients with cancer and related diseases. Much hope in research and healthcare in general is depending on such register-based analyses in order to comprehensively consider the features of a highly diverse population. Next to the data collection the cancer registries are responsible for data protection. To fulfill legal regulations, access to data has to be controlled in a strict way leading to sometimes bureaucratic and slow processes. The situation is especially complicated in Germany, since cancer data is distributed over numerous federal cancer registries. If a nationwide data evaluation is conducted a research team has to negotiate a separate contract with each cancer registry. In a joint work in progress effort of cancer registries, technical, medical, and economical experts we propose a different solution for cooperative data processing. Our approach aims for combining data in a virtual pool based on the selection criteria of individual requests from researchers. To achieve our goal, we adapt the Fraunhofer Medical Data Space as enabling technology. The architecture we propose will allow us to pool data of multiple partners regulated by data access policies. In doing so, each of the data sources can introduce its own rules and specifications on how data is used. Additionally, we add a digital consent management that will allow individual patients to decide how their data is used. Finally, we show the high potential of the cooperative analysis of distributed cancer data supported by the proposed solution in our approach.
-
PublicationCurrent Practices, Challenges, and Design Implications for Collaborative AR/VR Application Development( 2021)Augmented/Virtual Reality (AR/VR) is still a fragmented space to de- sign for due to the rapidly evolving hardware, the interdisciplinarity of teams, and a lack of standards and best practices. We interviewed 26 professional AR/VR designers and developers to shed light on their tasks, approaches, tools, and challenges. Based on their work and the artifacts they generated, we found that AR/VR application creators fulfill four roles: concept developers, interaction designers, content authors, and technical developers. One person often incorporates multiple roles and faces a variety of challenges during the design process from the initial contextual analysis to the deployment. From analysis of their tool sets, methods, and artifacts, we describe critical key challenges. Finally, we discus s the importance of prototyping for the communication in AR/VR development teams and highlight design implications for future tools to create a more usable AR/VR tool chain.
-
PublicationVorgehensmodell und Maßnahmenkatalog zur Entwicklung und Durchführung von beruflichen Weiterbildungen( 2021)
