MWMartin Welk

Martin Welk

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Martin Welk has a total of 39 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • XSRD-Net: EXplainable Stroke Relapse Detection

    by Christian Gapp, Elias Tappeiner, Martin Welk, Karl Fritscher, Stephanie Mangesius, Constantin Eisenschink, Philipp Deisl, Michael Knoflach, Astrid E. Grams, Elke R. Gizewski, et al.

  • Morphological PDEs with Rotationally Invariant Space-Fractional Derivatives

    by Martin Welk, Andreas Kleefeld, Michael Breuß, Bernhard Burgeth

  • Open Access

    What are you looking at? Modality contribution in multimodal medical deep learning

    by Christian Gapp, Elias Tappeiner, Martin Welk, Karl Fritscher, Elke R. Gizewski, Rainer Schubert

    Purpose: High dimensional, multimodal data can nowadays be analyzed by huge deep neural networks with little effort. Several fusion methods for bringing together different modalities have been developed. Given the prevalence of high-dimensional, multimodal patient data in medicine, the development of multimodal models marks a significant advancement. However, how these models process information from individual sources in detail is still underexplored. Methods: To this end, we implemented an occlusion-based modality contribution method that is both model- and performance agnostic. This method quantitatively measures the importance of each modality in the dataset for the model to fulfill its task. We applied our method to three different multimodal medical problems for experimental purposes. Results: Herein we found that some networks have modality preferences that tend to unimodal collapses, while some datasets are imbalanced from the ground up. Moreover, we provide fine-grained quantitative and visual attribute importance for each modality. Conclusion: Our metric offers valuable insights that can support the advancement of multimodal model development and dataset creation. By introducing this method, we contribute to the growing field of interpretability in deep learning for multimodal research. This approach helps to facilitate the integration of multimodal AI into clinical practice. Our code is publicly available at https://github.com/ChristianGappGit/MC_MMD


  • Open Access

    Matrix-Valued LogSumExp Approximation for Colour Morphology

    by Marvin Kahra, Michael Breuß, Andreas Kleefeld, Martin Welk

    Mathematical morphology is a part of image processing that employs a moving window to modify pixel values through the application of specific operations. The supremum and infimum are pivotal concepts, yet defining them in a general sense for high-dimensional data such as colour is a challenging endeavour. As a result, a number of different approaches have been taken to try to find a solution, with certain compromises being made along the way. In this paper, we present an analysis of a novel approach that replaces the supremum within a morphological operation with the LogExp approximation of the maximum for matrix-valued colours. This approach has the advantage of extending the associativity of dilation from the one-dimensional to the higher-dimensional case. Furthermore, the minimality property is investigated and a relaxation specified to ensure that the approach is continuously dependent on the input data.


  • Open Access

    Head and Neck Tumor Segmentation on MRIs with Fast and Resource-Efficient Staged nnU-Nets

    by Elias Tappeiner, Christian Gapp, Martin Welk, Rainer Schubert

    MRI-guided radiotherapy (RT) planning offers key advantages over conventional CT-based methods, including superior soft tissue contrast and the potential for daily adaptive RT due to the reduction of the radiation burden. In the Head and Neck (HN) region labor-intensive and time-consuming tumor segmentation still limits full utilization of MRI-guided adaptive RT. The HN Tumor Segmentation for MR-Guided Applications 2024 challenge (HNTS-MRG) aims to improve automatic tumor segmentation on MRI images by providing a dataset with reference annotations for the tasks of pre-RT and mid-RT planning. In this work, we present our approach for the HNTS-MRG challenge. Based on the insights of a thorough literature review we implemented a fast and resource-efficient two-stage segmentation method using the nnU-Net architecture with residual encoders as a backbone. In our two-stage approach we use the segmentation results of a first training round to guide the sampling process for a second refinement stage. For the pre-RT task, we achieved competitive results using only the first-stage nnU-Net. For the mid-RT task, we could significantly increase the segmentation performance of the basic first stage nnU-Net by utilizing the prior knowledge of the pre-RT plan as an additional input for the second stage refinement network. As team alpinists we achieved an aggregated Dice Coefficient of 80.97 for the pre-RT and 69.84 for the mid-RT task on the online test set of the challenge. Our code and trained model weights for the two-stage nnU-Net approach with residual encoders are available at https://github.com/elitap/hntsmrg24.


  • Open Access

    Multimodal Medical Disease Classification with LLaMA II

    by Christian Gapp, Elias Tappeiner, Martin Welk, Rainer Schubert

    Medical patient data is always multimodal. Images, text, age, gender, histopathological data are only few examples for different modalities in this context. Processing and integrating this multimodal data with deep learning based methods is of utmost interest due to its huge potential for medical procedure such as diagnosis and patient treatment planning. In this work we retrain a multimodal transformer-based model for disease classification. To this end we use the text-image pair dataset from OpenI consisting of 2D chest X-rays associated with clinical reports. Our focus is on fusion methods for merging text and vision information extracted from medical datasets. Different architecture structures with a LLaMA II backbone model are tested. Early fusion of modality specific features creates better results with the best model reaching 97.10% mean AUC than late fusion from a deeper level of the architecture (best model: 96.67% mean AUC). Both outperform former classification models tested on the same multimodal dataset. The newly introduced multimodal architecture can be applied to other multimodal datasets with little effort and can be easily adapted for further research, especially, but not limited to, the field of medical AI.


  • Open Access

    Matrix-Valued LogSumExp Approximation for Colour Morphology

    by Marvin Kahra, Michael Breuß, Andreas Kleefeld, Martin Welk

    Mathematical morphology is a part of image processing that uses a window that moves across the image to change certain pixels according to certain operations. The concepts of supremum and infimum play a crucial role here, but it proves challenging to define them generally for higher-dimensional data, such as colour representations. Numerous approaches have therefore been taken to solve this problem with certain compromises. In this paper we will analyse the construction of a new approach, which we have already presented experimentally in paper [Kahra, M., Breuß, M., Kleefeld, A., Welk, M., DGMM 2024, pp. 325-337]. This is based on a method by Burgeth and Kleefeld [Burgeth, B., Kleefeld, A., ISMM 2013, pp. 243-254], who regard the colours as symmetric $2\times2$ matrices and compare them by means of the Loewner order in a bi-cone through different suprema. However, we will replace the supremum with the LogExp approximation for the maximum instead. This allows us to transfer the associativity of the dilation from the one-dimensional case to the higher-dimensional case. In addition, we will investigate the minimality property and specify a relaxation to ensure that our approach is continuously dependent on the input data.


  • Open Access

    An Approach to Colour Morphological Supremum Formation using the LogSumExp Approximation

    by Marvin Kahra, Michael Breuß, Andreas Kleefeld, Martin Welk

    Mathematical morphology is a part of image processing that has proven to be fruitful for numerous applications. Two main operations in mathematical morphology are dilation and erosion. These are based on the construction of a supremum or infimum with respect to an order over the tonal range in a certain section of the image. The tonal ordering can easily be realised in grey-scale morphology, and some morphological methods have been proposed for colour morphology. However, all of these have certain limitations. In this paper we present a novel approach to colour morphology extending upon previous work in the field based on the Loewner order. We propose to consider an approximation of the supremum by means of a log-sum exponentiation introduced by Maslov. We apply this to the embedding of an RGB image in a field of symmetric $2\times2$ matrices. In this way we obtain nearly isotropic matrices representing colours and the structural advantage of transitivity. In numerical experiments we highlight some remarkable properties of the proposed approach.