Sessions

More than 40% of the adult population suffers from functional gastrointestinal disorders, now considered disorders of the “gut-brain axis” (GBA) interactions, a very complex bidirectional neural, endocrine, immune, and humoral communication system modulated by the microbiota. To help discover, understand, and manage GBA disorders, the OnePlanet research center is developing digital twins focused on the GBA, combining novel sensors with artificial intelligence algorithms providing descriptive, diagnostic, predictive or prescriptive feed-back.

Deep learning models for radiology are typically deployed either through cloud-based platforms, through on-premises infrastructures, or though heavyweight viewers. This tends to restrict the audience of deep learning models to radiologists working in state-of-the-art hospitals, which raises concerns about the democratization of deep learning for medical imaging, most notably in the context of research and education. We show that complex deep learning models can be applied directly inside Web browsers, without resorting to any external computation infrastructure, thanks to the use of WebAssembly, and we release our code as free and open-source software. This opens the path to the use of teleradiology solutions as an effective way to distribute, teach, and evaluate deep learning architectures.

We benchmark the influence of four types of noise annotated in the public dataset PTB-XL on the accuracy of a Deep Learning-based method for atrial fibrillation detection in 12-lead ECGs. For this we use the noise metadata provided by human experts as well as a quantitative signal-to-noise ratio for assigning a signal quality to each ECG and analyze the respective DL accuracy. We conclude that the issue of processing noisy electrocardiography data can be addressed successfully by Deep Learning methods.

Age-related macular degeneration (AMD) is a major cause of blindness in the West. This study utilizes spectral domain optical coherence tomography (SD-OCT) for retinal imaging and deep learning for analysis. A convolutional neural network (CNN) was trained on 1300 expert-annotated SD-OCT scans to identify AMD biomarkers. Transfer learning improved the CNN's performance using weights from a separate classifier. The model accurately detects and segments AMD biomarkers in OCT scans, potentially assisting in patient prioritization and reducing ophthalmologists' workloads.

Separate registration required: https://www.mie2023.org/tutorials, The open-source Python framework AUCMEDI offers a solution to the describedchallenges. The software package not only offers a library as a 'high-level' API for thestandardized construction of modern medical image classification pipelines, but alsoreproducible installation and direct application via Dockerization and automatichyperparameter detection. With AUCMEDI, researchers are able to set up a completeas well as easy-to-integrate medical image classification pipeline with just a few linesof code. AUCMEDI is available as a Python package via PyPI ('pip install aucmedi')and as a repository via GitHub with detailed documentation, examples, and bindings tomodern DevOps (CI/CD) techniques: https://frankkramer-lab.github.io/aucmedi/.