Filters

2015 - 201912020 - 20231

More filters

2022 - 20242
Reset filters

Settings

Publisher: [London] : Nature Publishing Group UK × Subject: evolution ×

Search Results

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Evolutionary design of explainable algorithms for biomedical image segmentation
    ([London] : Nature Publishing Group UK, 2023) Cortacero, Kévin; McKenzie, Brienne; Müller, Sabina; Khazen, Roxana; Lafouresse, Fanny; Corsaut, Gaëlle; Van Acker, Nathalie; Frenois, François-Xavier; Lamant, Laurence; Meyer, Nicolas; Vergier, Béatrice; Wilson, Dennis G.; Luga, Hervé; Staufer, Oskar; Dustin, Michael L.; Valitutti, Salvatore; Cussat-Blanc, Sylvain
    An unresolved issue in contemporary biomedicine is the overwhelming number and diversity of complex images that require annotation, analysis and interpretation. Recent advances in Deep Learning have revolutionized the field of computer vision, creating algorithms that compete with human experts in image segmentation tasks. However, these frameworks require large human-annotated datasets for training and the resulting “black box” models are difficult to interpret. In this study, we introduce Kartezio, a modular Cartesian Genetic Programming-based computational strategy that generates fully transparent and easily interpretable image processing pipelines by iteratively assembling and parameterizing computer vision functions. The pipelines thus generated exhibit comparable precision to state-of-the-art Deep Learning approaches on instance segmentation tasks, while requiring drastically smaller training datasets. This Few-Shot Learning method confers tremendous flexibility, speed, and functionality to this approach. We then deploy Kartezio to solve a series of semantic and instance segmentation problems, and demonstrate its utility across diverse images ranging from multiplexed tissue histopathology images to high resolution microscopy images. While the flexibility, robustness and practical utility of Kartezio make this fully explicable evolutionary designer a potential game-changer in the field of biomedical image processing, Kartezio remains complementary and potentially auxiliary to mainstream Deep Learning approaches.
  • Thumbnail Image
    Item
    Retrieving spin textures on curved magnetic thin films with full-field soft X-ray microscopies
    ([London] : Nature Publishing Group UK, 2015) Streubel, Robert; Kronast, Florian; Fischer, Peter; Parkinson, Dula; Schmidt, Oliver G.; Makarov, Denys
    X-ray tomography is a well-established technique to characterize 3D structures in material sciences and biology; its magnetic analogue--magnetic X-ray tomography--is yet to be developed. Here we demonstrate the visualization and reconstruction of magnetic domain structures in a 3D curved magnetic thin films with tubular shape by means of full-field soft X-ray microscopies. The 3D arrangement of the magnetization is retrieved from a set of 2D projections by analysing the evolution of the magnetic contrast with varying projection angle. Using reconstruction algorithms to analyse the angular evolution of 2D projections provides quantitative information about domain patterns and magnetic coupling phenomena between windings of azimuthally and radially magnetized tubular objects. The present approach represents a first milestone towards visualizing magnetization textures of 3D curved thin films with virtually arbitrary shape.