@misc{OPUS4-8974, title = {Bildverarbeitung f{\"u}r die Medizin 2026 : Proceedings, German Conference on Medical Image Computing, L{\"u}beck March 15-17, 2026}, editor = {Handels, Heinz and Breininger, Katharina and Deserno, Thomas M. and Maier, Andreas and Maier-Hein, Klaus H. and Palm, Christoph and Tolxdorff, Thomas}, publisher = {Springer Vieweg}, address = {Wiesbaden}, isbn = {978-3-658-51099-2}, issn = {1431-472X}, doi = {10.1007/978-3-658-51100-5}, pages = {xxix, 496}, abstract = {Die Konferenz "BVM - Bildverarbeitung f{\"u}r die Medizin" ist seit vielen Jahren als die nationale Plattform f{\"u}r den Austausch von Ideen und die Diskussion der neuesten Forschungsergebnisse im Bereich der Medizinischen Bildverarbeitung und der K{\"u}nstlichen Intelligenz (KI) etabliert. Auch 2026 haben (junge) Wissenschaftler*innen, Industrie und Anwender*innen diesen Austausch vertieft. Die Beitr{\"a}ge dieses Bandes - die meisten davon in englischer Sprache - umfassen alle Bereiche der medizinischen Bildverarbeitung, insbesondere die Bildgebung und -akquisition, Segmentierung und Analyse, Registrierung, Visualisierung und Animation, computerunterst{\"u}tzte Diagnose sowie bildgest{\"u}tzte Therapieplanung und Therapie. Hierbei kommen Methoden des maschinellen Lernens, der biomechanischen Modellierung sowie der Validierung und Qualit{\"a}tssicherung zum Einsatz.}, subject = {Bildverarbeitung}, language = {de} } @phdthesis{SouzaJr, author = {Souza Jr., Luis Antonio de}, title = {Computer-assisted diagnosis of Barrett's esophagus using machine learning techniques}, publisher = {Universidade Federal de S{\~a}o Carlos}, pages = {262}, abstract = {Esophageal adenocarcinoma is an illness that is usually hard to detect at the early stages in the presence of Barrett's esohagus. The development of automatic evaluation systems of such illness may be very useful, thus assisting the experts in the neoplastic region detection. With the strong growth of machine learning techniques aiming to improve the effectivess of medical diagnosis, the use of such approaches characterizes a strong scenario to be explored for the early diagnosis of esophageal adenocarcinoma. Barrett's esophagus as a predecessor of adenocarcinoma can be explained by some risk factors, such as obesity, smoking, and late medical diagnosis. This project proposes the development of new computer vision and machine learning techniques to assist the automatic diagnosis of the esophageal adenocarcinioma based on the evaluation of two kind of features: (i) handcrafted features, calculated by means of human knowledge using some image processing technique and; (ii) deeply-learnable features, calculated exclusively based on deep learning techniques. From the extensive application of global and local protocols for the models proposed in this work, the description of cancer-affected images and Barrett's esophagus-affected samples were generalized and deeply evaluated using, for example, classifiers such as Support Vector Machines, ResNet-50 and the combination of descriptions by handcrafted and deeply-learnable features. Also, the behavior of the automatic definition of key-points within the evaluated techniques was observed, something of a paramount importance nowadays to guarantee transparency and reliability in the decisions made by computational techniques. Thus, this project contributes to both the computational and medical fields, introducing new classifiers, approaches and interpretation of the class generalization process, in addition to proposing fast and precise manners to define cancer, delivering important and novel results concerning the accurate identification of cancer in samples affected by Barrett's esophagus, showing values around 95\% of correct identification rates and arranged in a collection of scientific works developed by the author during the research period and submitted/published to date.}, language = {en} } @misc{GutbrodRauberWeberNunesetal., author = {Gutbrod, Max and Rauber, David and Weber Nunes, Danilo and Palm, Christoph}, title = {Cropped single instrument frames subset from Cholec80 [Data set]}, doi = {10.5281/zenodo.14921670}, abstract = {This dataset is a subset of the original Cholec80 dataset and is used by the OpenMIBOOD framework to evaluate a specific out-of-distribution setting. When using this dataset, it is mandatory to cite the corresponding publication (OpenMIBOOD) and to follow the acknowledgement and citation requirements of the original dataset (Cholec80). The original Cholec80 dataset (associated paper,Homepage) consists of 80 cholecystectomy surgery videos recorded at 25 fps, performed by 13 surgeons. It includes phase annotations (25 fps) and tool presence labels (1 fps), with phase definitions provided by a senior surgeon. A tool is considered present if at least half of its tip is visible. The dataset categorizes tools into seven types: Grasper, Bipolar, Hook, Scissors, Clipper, Irrigator, and Specimen bag. Multiple tools may be present in each frame. Additionally, 76 of the 80 videos exhibit a strong black vignette. For this dataset subset, frames were extracted based on tool presence labels, selecting only those containing Grasper, Bipolar, Hook, or Clipper while ensuring that only a single tool appears per frame. To enhance visual consistency, the black vignette was removed by extracting an inner rectangular region, where applicable.}, language = {en} } @unpublished{LaubmannReschke, author = {Laubmann, Julia and Reschke, Johannes}, title = {Tackling fake images in cybersecurity - interpretation of a StyleGAN and lifting its black-box}, doi = {10.48550/arXiv.2507.13722}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:898-opus4-83930}, pages = {11}, abstract = {In today's digital age, concerns about the dangers of AI-generated images are increasingly common. One powerful tool in this domain is StyleGAN (style-based generative adversarial networks), a generative adversarial network capable of producing highly realistic synthetic faces. To gain a deeper understanding of how such a model operates, this work focuses on analyzing the inner workings of StyleGAN's generator component. Key architectural elements and techniques, such as the Equalized Learning Rate, are explored in detail to shed light on the model's behavior. A StyleGAN model is trained using the PyTorch framework, enabling direct inspection of its learned weights. Through pruning, it is revealed that a significant number of these weights can be removed without drastically affecting the output, leading to reduced computational requirements. Moreover, the role of the latent vector -- which heavily influences the appearance of the generated faces -- is closely examined. Global alterations to this vector primarily affect aspects like color tones, while targeted changes to individual dimensions allow for precise manipulation of specific facial features. This ability to finetune visual traits is not only of academic interest but also highlights a serious ethical concern: the potential misuse of such technology. Malicious actors could exploit this capability to fabricate convincing fake identities, posing significant risks in the context of digital deception and cybercrime.}, language = {en} } @misc{OPUS4-7967, title = {Bildverarbeitung f{\"u}r die Medizin 2025}, editor = {Palm, Christoph and Breininger, Katharina and Deserno, Thomas M. and Handels, Heinz and Maier, Andreas and Maier-Hein, Klaus H. and Tolxdorff, Thomas M.}, publisher = {Springer Fachmedien Wiesbaden}, address = {Wiesbaden}, isbn = {978-3-658-47421-8}, issn = {1431-472X}, doi = {10.1007/978-3-658-47422-5}, pages = {XXIII, 354}, abstract = {Die Konferenz "BVM - Bildverarbeitung f{\"u}r die Medizin" ist seit vielen Jahren als die nationale Plattform f{\"u}r den Austausch von Ideen und die Diskussion der neuesten Forschungsergebnisse im Bereich der Medizinischen Bildverarbeitung und der K{\"u}nstlichen Intelligenz (KI) etabliert. Auch 2025 werden wir aktuelle Forschungsergebnisse vorstellen und Gespr{\"a}che zwischen (jungen) Wissenschaftler*innen, Industrie und Anwender*innen vertiefen. Die Beitr{\"a}ge dieses Bandes - die meisten davon in englischer Sprache - umfassen alle Bereiche der medizinischen Bildverarbeitung, insbesondere die Bildgebung und -akquisition, Segmentierung und Analyse, Registrierung, Visualisierung und Animation, computerunterst{\"u}tzte Diagnose sowie bildgest{\"u}tzte Therapieplanung und Therapie. Hierbei kommen Methoden des maschinellen Lernens, der biomechanischen Modellierung sowie der Validierung und Qualit{\"a}tssicherung zum Einsatz. Das Kapitel "Leveraging multiple total body segmentators and anatomy-informed post-processing for segmenting bones in Lung CTs" ist unter einer Creative Commons Attribution 4.0 International License {\"u}ber link.springer.com frei verf{\"u}gbar (Open Access). Die Herausgebenden Prof. Palm forscht im Bereich KI f{\"u}r die Medizin mit einem Schwerpunkt in der Analyse endoskopischer Bilddaten zur computerunterst{\"u}tzten Diagnose und Therapie. Prof. Breininger entwickelt robuste Ans{\"a}tze des maschinellen Lernens in verschiedenen interdisziplin{\"a}ren Bereichen, mit einem Schwerpunkt auf medizinischen Bilddaten. Prof. Deserno forscht in Biosignal- und Bilderzeugung und -verarbeitung, insbesondere in der videobasierten Vitaldatenmessung. Prof. Handels entwickelt problemoptimierte, lernf{\"a}hige Bildverarbeitungsmethoden und integriert diese in hybride Bildverarbeitungssysteme zur Unterst{\"u}tzung der medizinischen Diagnostik und Therapie. Prof. Maier entwickelt Anwendungen in der medizinischen Bildverarbeitung zur Diagnoseunterst{\"u}tzung bis hin zur Schichtbildberechnung durch k{\"u}nstliche Intelligenz. Prof. Maier-Hein forscht im Bereich maschinelles Lernen und entwickelt Open-Source-L{\"o}sungen wie das Medical Imaging Interaction Toolkit (MITK), Kaapana oder das nnU-Net. Prof. em. Tolxdorff ist Experte f{\"u}r maschinelles Lernen, biomedizinisches Datenmanagement, Datenvisualisierung und -analyse sowie Medizinproduktentwicklung in klinischen Workflows.}, subject = {Bildverarbeitung}, language = {de} } @article{BrollGoldhackerHahneletal., author = {Broll, Alexander and Goldhacker, Markus and Hahnel, Sebastian and Rosentritt, Martin}, title = {Generative deep learning approaches for the design of dental restorations: A narrative review}, series = {Journal of Dentistry}, volume = {145}, journal = {Journal of Dentistry}, publisher = {Elsevier}, issn = {0300-5712}, doi = {10.1016/j.jdent.2024.104988}, pages = {10}, abstract = {Objectives: This study aims to explore and discuss recent advancements in tooth reconstruction utilizing deep learning (DL) techniques. A review on new DL methodologies in partial and full tooth reconstruction is conducted. Data/Sources: PubMed, Google Scholar, and IEEE Xplore databases were searched for articles from 2003 to 2023. Study selection: The review includes 9 articles published from 2018 to 2023. The selected articles showcase novel DL approaches for tooth reconstruction, while those concentrating solely on the application or review of DL methods are excluded. The review shows that data is acquired via intraoral scans or laboratory scans of dental plaster models. Common data representations are depth maps, point clouds, and voxelized point clouds. Reconstructions focus on single teeth, using data from adjacent teeth or the entire jaw. Some articles include antagonist teeth data and features like occlusal grooves and gap distance. Primary network architectures include Generative Adversarial Networks (GANs) and Transformers. Compared to conventional digital methods, DL-based tooth reconstruction reports error rates approximately two times lower. Conclusions: Generative DL models analyze dental datasets to reconstruct missing teeth by extracting insights into patterns and structures. Through specialized application, these models reconstruct morphologically and functionally sound dental structures, leveraging information from the existing teeth. The reported advancements facilitate the feasibility of DL-based dental crown reconstruction. Beyond GANs and Transformers with point clouds or voxels, recent studies indicate promising outcomes with diffusion-based architectures and innovative data representations like wavelets for 3D shape completion and inference problems. Clinical significance: Generative network architectures employed in the analysis and reconstruction of dental structures demonstrate notable proficiency. The enhanced accuracy and efficiency of DL-based frameworks hold the potential to enhance clinical outcomes and increase patient satisfaction. The reduced reconstruction times and diminished requirement for manual intervention may lead to cost savings and improved accessibility of dental services.}, language = {en} } @unpublished{MendelRueckertWilhelmetal., author = {Mendel, Robert and R{\"u}ckert, Tobias and Wilhelm, Dirk and R{\"u}ckert, Daniel and Palm, Christoph}, title = {Motion-Corrected Moving Average: Including Post-Hoc Temporal Information for Improved Video Segmentation}, doi = {10.48550/arXiv.2403.03120}, pages = {9}, abstract = {Real-time computational speed and a high degree of precision are requirements for computer-assisted interventions. Applying a segmentation network to a medical video processing task can introduce significant inter-frame prediction noise. Existing approaches can reduce inconsistencies by including temporal information but often impose requirements on the architecture or dataset. This paper proposes a method to include temporal information in any segmentation model and, thus, a technique to improve video segmentation performance without alterations during training or additional labeling. With Motion-Corrected Moving Average, we refine the exponential moving average between the current and previous predictions. Using optical flow to estimate the movement between consecutive frames, we can shift the prior term in the moving-average calculation to align with the geometry of the current frame. The optical flow calculation does not require the output of the model and can therefore be performed in parallel, leading to no significant runtime penalty for our approach. We evaluate our approach on two publicly available segmentation datasets and two proprietary endoscopic datasets and show improvements over a baseline approach.}, subject = {Deep Learning}, language = {en} } @article{SouzaJrPachecoPassosetal., author = {Souza Jr., Luis Antonio de and Pacheco, Andr{\´e} G.C. and Passos, Leandro A. and Santana, Marcos Cleison S. and Mendel, Robert and Ebigbo, Alanna and Probst, Andreas and Messmann, Helmut and Palm, Christoph and Papa, Jo{\~a}o Paulo}, title = {DeepCraftFuse: visual and deeply-learnable features work better together for esophageal cancer detection in patients with Barrett's esophagus}, series = {Neural Computing and Applications}, volume = {36}, journal = {Neural Computing and Applications}, publisher = {Springer}, address = {London}, doi = {10.1007/s00521-024-09615-z}, pages = {10445 -- 10459}, abstract = {Limitations in computer-assisted diagnosis include lack of labeled data and inability to model the relation between what experts see and what computers learn. Even though artificial intelligence and machine learning have demonstrated remarkable performances in medical image computing, their accountability and transparency level must be improved to transfer this success into clinical practice. The reliability of machine learning decisions must be explained and interpreted, especially for supporting the medical diagnosis. While deep learning techniques are broad so that unseen information might help learn patterns of interest, human insights to describe objects of interest help in decision-making. This paper proposes a novel approach, DeepCraftFuse, to address the challenge of combining information provided by deep networks with visual-based features to significantly enhance the correct identification of cancerous tissues in patients affected with Barrett's esophagus (BE). We demonstrate that DeepCraftFuse outperforms state-of-the-art techniques on private and public datasets, reaching results of around 95\% when distinguishing patients affected by BE that is either positive or negative to esophageal cancer.}, subject = {Deep Learning}, language = {en} } @misc{OPUS4-7115, title = {Bildverarbeitung f{\"u}r die Medizin 2024}, editor = {Maier, Andreas and Deserno, Thomas M. and Handels, Heinz and Maier-Hein, Klaus H. and Palm, Christoph and Tolxdorff, Thomas}, publisher = {Springer}, address = {Wiesbaden}, isbn = {978-3-658-44037-4}, issn = {1431-472X}, doi = {10.1007/978-3-658-44037-4}, pages = {370}, abstract = {Seit mehr als 25 Jahren ist der Workshop "Bildverarbeitung f{\"u}r die Medizin" als erfolgreiche Veranstaltung etabliert. Ziel ist auch 2024 wieder die Darstellung aktueller Forschungsergebnisse und die Vertiefung der Gespr{\"a}che zwischen Wissenschaftlern, Industrie und Anwendern. Die Beitr{\"a}ge dieses Bandes - viele davon in englischer Sprache - umfassen alle Bereiche der medizinischen Bildverarbeitung, insbesondere die Bildgebung und -akquisition, Segmentierung und Analyse, Visualisierung und Animation, computerunterst{\"u}tzte Diagnose sowie bildgest{\"u}tzte Therapieplanung und Therapie. Hierbei kommen Methoden des maschinelles Lernens, der biomechanischen Modellierung sowie der Validierung und Qualit{\"a}tssicherung zum Einsatz.}, subject = {Bildverarbeitung}, language = {de} } @article{RueckertRueckertPalm, author = {R{\"u}ckert, Tobias and R{\"u}ckert, Daniel and Palm, Christoph}, title = {Methods and datasets for segmentation of minimally invasive surgical instruments in endoscopic images and videos: A review of the state of the art}, series = {Computers in Biology and Medicine}, volume = {169}, journal = {Computers in Biology and Medicine}, publisher = {Elsevier}, address = {Amsterdam}, doi = {10.1016/j.compbiomed.2024.107929}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:898-opus4-69830}, pages = {24}, abstract = {In the field of computer- and robot-assisted minimally invasive surgery, enormous progress has been made in recent years based on the recognition of surgical instruments in endoscopic images and videos. In particular, the determination of the position and type of instruments is of great interest. Current work involves both spatial and temporal information, with the idea that predicting the movement of surgical tools over time may improve the quality of the final segmentations. The provision of publicly available datasets has recently encouraged the development of new methods, mainly based on deep learning. In this review, we identify and characterize datasets used for method development and evaluation and quantify their frequency of use in the literature. We further present an overview of the current state of research regarding the segmentation and tracking of minimally invasive surgical instruments in endoscopic images and videos. The paper focuses on methods that work purely visually, without markers of any kind attached to the instruments, considering both single-frame semantic and instance segmentation approaches, as well as those that incorporate temporal information. The publications analyzed were identified through the platforms Google Scholar, Web of Science, and PubMed. The search terms used were "instrument segmentation", "instrument tracking", "surgical tool segmentation", and "surgical tool tracking", resulting in a total of 741 articles published between 01/2015 and 07/2023, of which 123 were included using systematic selection criteria. A discussion of the reviewed literature is provided, highlighting existing shortcomings and emphasizing the available potential for future developments.}, subject = {Deep Learning}, language = {en} } @misc{OPUS4-6079, title = {Bildverarbeitung f{\"u}r die Medizin 2023}, editor = {Deserno, Thomas M. and Handels, Heinz and Maier, Andreas and Maier-Hein, Klaus H. and Palm, Christoph and Tolxdorff, Thomas}, publisher = {Springer Vieweg}, address = {Wiesbaden}, isbn = {978-3-658-41656-0}, issn = {1431-472X}, doi = {10.1007/978-3-658-41657-7}, pages = {317}, abstract = {Seit mehr als 25 Jahren ist der Workshop "Bildverarbeitung f{\"u}r die Medizin" als erfolgreiche Veranstaltung etabliert. Ziel ist auch 2023 wieder die Darstellung aktueller Forschungsergebnisse und die Vertiefung der Gespr{\"a}che zwischen Wissenschaftlern, Industrie und Anwendern. Die Beitr{\"a}ge dieses Bandes - viele davon in englischer Sprache - umfassen alle Bereiche der medizinischen Bildverarbeitung, insbesondere die Bildgebung und -akquisition, Segmentierung und Analyse, Visualisierung und Animation, computerunterst{\"u}tzte Diagnose sowie bildgest{\"u}tzte Therapieplanung und Therapie. Hierbei kommen Methoden des maschinelles Lernens, der biomechanischen Modellierung sowie der Validierung und Qualit{\"a}tssicherung zum Einsatz.}, subject = {Bildverarbeitung}, language = {de} } @inproceedings{BirkenmaierKrenkel, author = {Birkenmaier, Clemens and Krenkel, Lars}, title = {Convolutional Neural Networks for Approximation of Internal Non-Newtonian Multiphase Flow Fields}, series = {14th World Congress on Computational Mechanics (WCCM), ECCOMAS Congress 2020: 19-24 July 2020, Paris, France}, booktitle = {14th World Congress on Computational Mechanics (WCCM), ECCOMAS Congress 2020: 19-24 July 2020, Paris, France}, editor = {Chinesta, F. and Abgrall, R. and Allix, O. and Kalistke, M}, publisher = {CIMNE}, doi = {10.23967/wccm-eccomas.2020.107}, abstract = {Neural networks (NNs) as an alternative method for universal approximation of differential equations have proven to be computationally efficient and still sufficiently accurate compared to established methods such as the finite volume method (FVM). Additionally, analysing weights and biases can give insights into the underlying physical laws. FVM and NNs are both based upon spacial discretisation. Since a Cartesian and equidistant grid is a raster graphics, image-to-image regression techniques can be used to predict phase velocity fields as well as particle and pressure distributions from simple mass flow boundary conditions. The impact of convolution layer depth and number of channels of a ConvolutionDeconvolution Regression Network (CDRN), on prediction performance of internal non-Newtownian multiphase flows is investigated. Parametric training data with 2055 sets is computed using FVM. To capture significant non-Newtownian effects of a particle-laden fluid (e.g. blood) flowing through small and non-straight channels, an Euler-Euler multiphase approach is used. The FVM results are normalized and mapped onto an equidistant grid as supervised learning target. The investigated NNs consist of n= {3, 5, 7} corresponding encoding/decoding blocks and different skip connections. Regardless of the convolution depth (i.e. number of blocks), the deepest spacial down-sampling via strided convolution is adjusted to result in a 1 × 1 × f · 2nfeature map, with f = {8, 16, 32}. The prediction performance expressed is as channel-averaged normalized root mean squared error (NRMSE). With a NRMSE of < 2 · 10-3, the best preforming NN has f = 32 initial feature maps, a kernel size of k = 4, n = 5 blocks and dense skip connections. Average inference time from this NN takes < 7 · 10-3s. Worst accuracy at NRMSE of approx 9 · 10-3is achieved without any skips, at k = 2, f = 16 and n = 3, but deployment takes only < 2 · 10-3s Given an adequate training, the prediction accuracy improves with convolution depth, where more features have higher impact on deeper NNs. Due to skip connections and batch normalisation, training is similarly efficient, regardless of the depth. This is further improved by blocks with dense connections, but at the price of a drastically larger model. Depending on geometrical complexity, spacial resolution is critical, as it increases the number of learnables and memory requirements massively.}, language = {en} } @misc{ScheppachMendelProbstetal., author = {Scheppach, Markus W. and Mendel, Robert and Probst, Andreas and Meinikheim, Michael and Palm, Christoph and Messmann, Helmut and Ebigbo, Alanna}, title = {Intraprozedurale Strukturerkennung bei Third-Space Endoskopie mithilfe eines Deep-Learning Algorithmus}, series = {Zeitschrift f{\"u}r Gastroenterologie}, volume = {60}, journal = {Zeitschrift f{\"u}r Gastroenterologie}, number = {04}, publisher = {Thieme}, address = {Stuttgart}, doi = {10.1055/s-0042-1745652}, pages = {e250-e251}, abstract = {Einleitung Third-Space Interventionen wie die endoskopische Submukosadissektion (ESD) und die perorale endoskopische Myotomie (POEM) sind technisch anspruchsvoll und mit einem erh{\"o}hten Risiko f{\"u}r intraprozedurale Komplikationen wie Blutung oder Perforation assoziiert. Moderne Computerprogramme zur Unterst{\"u}tzung bei diagnostischen Entscheidungen werden unter Einsatz von k{\"u}nstlicher Intelligenz (KI) in der Endoskopie bereits erfolgreich eingesetzt. Ziel der vorliegenden Arbeit war es, relevante anatomische Strukturen mithilfe eines Deep-Learning Algorithmus zu detektieren und segmentieren, um die Sicherheit und Anwendbarkeit von ESD und POEM zu erh{\"o}hen. Methoden Zw{\"o}lf Videoaufnahmen in voller L{\"a}nge von Third-Space Endoskopien wurden aus der Datenbank des Universit{\"a}tsklinikums Augsburg extrahiert. 1686 Einzelbilder wurden f{\"u}r die Kategorien Submukosa, Blutgef{\"a}ß, Dissektionsmesser und endoskopisches Instrument annotiert und segmentiert. Mit diesem Datensatz wurde ein DeepLabv3+neuronales Netzwerk auf der Basis eines ResNet mit 101 Schichten trainiert und intern anhand der Parameter Intersection over Union (IoU), Dice Score und Pixel Accuracy validiert. Die F{\"a}higkeit des Algorithmus zur Gef{\"a}ßdetektion wurde anhand von 24 Videoclips mit einer Spieldauer von 7 bis 46 Sekunden mit 33 vordefinierten Gef{\"a}ßen evaluiert. Anhand dieses Tests wurde auch die Gef{\"a}ßdetektionsrate eines Experten in der Third-Space Endoskopie ermittelt. Ergebnisse Der Algorithmus zeigte eine Gef{\"a}ßdetektionsrate von 93,94\% mit einer mittleren Rate an falsch positiven Signalen von 1,87 pro Minute. Die Gef{\"a}ßdetektionsrate des Experten lag bei 90,1\% ohne falsch positive Ergebnisse. In der internen Validierung an Einzelbildern wurde eine IoU von 63,47\%, ein mittlerer Dice Score von 76,18\% und eine Pixel Accuracy von 86,61\% ermittelt. Zusammenfassung Dies ist der erste KI-Algorithmus, der f{\"u}r den Einsatz in der therapeutischen Endoskopie entwickelt wurde. Pr{\"a}limin{\"a}re Ergebnisse deuten auf eine mit Experten vergleichbare Detektion von Gef{\"a}ßen w{\"a}hrend der Untersuchung hin. Weitere Untersuchungen sind n{\"o}tig, um die Leistung des Algorithmus im Vergleich zum Experten genauer zu eruieren sowie einen m{\"o}glichen klinischen Nutzen zu ermitteln.}, language = {de} } @inproceedings{RauberMendelScheppachetal., author = {Rauber, David and Mendel, Robert and Scheppach, Markus W. and Ebigbo, Alanna and Messmann, Helmut and Palm, Christoph}, title = {Analysis of Celiac Disease with Multimodal Deep Learning}, series = {Bildverarbeitung f{\"u}r die Medizin 2022: Proceedings, German Workshop on Medical Image Computing, Heidelberg, June 26-28, 2022}, booktitle = {Bildverarbeitung f{\"u}r die Medizin 2022: Proceedings, German Workshop on Medical Image Computing, Heidelberg, June 26-28, 2022}, publisher = {Springer Vieweg}, address = {Wiesbaden}, doi = {10.1007/978-3-658-36932-3_25}, pages = {115 -- 120}, abstract = {Celiac disease is an autoimmune disorder caused by gluten that results in an inflammatory response of the small intestine.We investigated whether celiac disease can be detected using endoscopic images through a deep learning approach. The results show that additional clinical parameters can improve the classification accuracy. In this work, we distinguished between healthy tissue and Marsh III, according to the Marsh score system. We first trained a baseline network to classify endoscopic images of the small bowel into these two classes and then augmented the approach with a multimodality component that took the antibody status into account.}, language = {en} }