Journal of microscopy最新文献

Microscopy and image analysis play a vital role in parasitology research; they are critical for identifying parasitic organisms and elucidating their complex life cycles. Despite major advancements in imaging and analysis, several challenges remain. These include the integration of interdisciplinary data; information derived from various model organisms; and data acquired from clinical research. In our view, artificial intelligence-with the latest advances in machine and deep learning-holds enormous potential to address many of these challenges. This review addresses how artificial intelligence, machine learning and deep learning have been used in the field of parasitology-mainly focused on Apicomplexan, Diplomonad, and Kinetoplastid groups. We explore how gaps in our understanding could be filled by AI in future parasitology research and diagnosis in the field. Moreover, it addresses challenges and limitations currently faced in implementing and expanding the use of artificial intelligence across biomedical fields. The necessary increased collaboration between biologists and computational scientists will facilitate understanding, development, and implementation of the latest advances for both scientific discovery and clinical impact. Current and future AI tools hold the potential to revolutionise parasitology and expand One Health principles.

Image noise is a fundamental problem in fluorescence microscopy analysis, especially in live cell imaging applications where the number of detected photons is limited due to low power of excitation lasers to prevent phototoxicity during extended imaging experiments. The noise increases measurement uncertainty and complicates further image processing routines such as deconvolution, object detection and segmentation. State-of-the-art denoisers are computationally expensive and require training using large datasets, which are not available in cases of typical biological imaging experiments with rather scarce and unlabelled data. Here, we show that a denoiser can be trained using a single image containing a cropped out object of interest, where we exploit the symmetry often present in biological structures at molecular scales. As only a single example is used during training, our method can be trained even with limited computational resources, obtaining competitive denoising performance compared to the state-of-the-art methods.

Recently developed secondary nanobodies or single-domain antibodies present a powerful tool for immunodetection. Unlike traditional antibodies, their monovalence enables pre-association with primary antibodies prior to sample staining, presenting a straightforward affinity-based antibody labelling solution. This not only simplifies and streamlines immunoassays, it also supports multiplexed techniques where conflicts in the species of the desired primary antibodies preclude standard indirect immunostaining. Despite these advantages, the use of secondary nanobodies remains sparse, due perhaps to a lack of evaluation on their suitability for assays requiring quantification and an assessment of optimal protocols for their use. Here, we propose a set of experiments spanning total internal reflection fluorescence and confocal microscopies that can be used to validate secondary nanobody binding, specificity, and their propensity for mis-targeted binding in multiplex assays. Using these tools, we analysed the binding properties of commercially available secondary nanobodies and outline optimised protocols for their robust use.

Persistent challenges in indium bump defect detection for infrared focal plane array (IRFPA) include low detection accuracy and high false-positive rates caused by subtle deformations from process variations, imaging distortions and scale variations in microscopic imaging. To address these shortcomings, this paper proposes a defect detection method based on the multi-scale Demons deformation field difference. A workflow comprising global coarse registration, local deformation field optimisation, and adaptive defect segmentation was established. Leveraging a multi-scale Bilateral Total Variation (BTV)-regularised Demons model, we progressively refined local details through hierarchical deformation field computation. This approach suppressed noise-induced distortions while preserving abrupt transitions, which helps to improve the saliency of local defect signals. The method shows improved robustness against deformation, noise, and scale variations compared to conventional template-matching algorithms. It is designed to meet the requirements for industrial inspection and provides a framework for weak-feature extraction and precise defect identification across multiple imaging modalities-including wide-field, bright-field confocal, and dark-field confocal microscopy.

Smart microscopy represents a paradigm shift in biological imaging, moving from passive observation tools to active collaborators in scientific inquiry. Enabled by advances in automation, computational power, and artificial intelligence, these systems are now capable of adaptive decision-making and real-time experimental control. Here, we introduce a theoretical framework that reconceptualises smart microscopy as a partner in scientific investigation. Central to our framework is the concept of the 'epistemic-empirical divide' in cellular investigation, describing the gap between what is observable (empirical domain) and what must be understood (epistemic domain). We propose six core design principles: epistemic-empirical awareness, hierarchical context integration, an evolution from detection to perception, adaptive measurement frameworks, narrative synthesis capabilities, and cross-contextual reasoning. Together, these principles guide a multi-agent architecture designed to align empirical observation with the goals of scientific understanding. Our framework provides a roadmap for building microscopy systems that go beyond automation to actively support hypothesis generation, insight discovery, and theory development, redefining the role of scientific instruments in the process of knowledge creation.

Human erythrocytes (red blood cells; RBCs) undergo spontaneous disassembly after several hours of exposure to n-butyl acetate (nBA). Images of the morphological changes were captured in time-lapse sequences using differential interference contrast (DIC) light microscopy. On exposure to less than 10 mM nBA dramatic events did not take place, but with ∼60 mM aqueous solutions of nBA, discocytes became spherical with a single contiguous 'geographical' indentation patch. Over the next ∼2 h the cells became echinocyte-like with rounded projections; and several hours later they discharged filaments that writhed in Brownian motion. In parallel with these changes, vesicles budded from the cells, followed by their alignment on the filaments, like balloons on a string. The vesicles then serially fused, finally giving rise to a single large vesicle that was ∼0.1-0.2 times the diameter of the spherical parent cell; it then fused with the parent cell that a short while later ruptured and became a 'ghost'. Owing to the striking nature of this phenomenon that was evocative of party decor, the term coined for it was Feierzeit (German: celebration time). The background to this serendipitous discovery, and the deductive odyssey that identified the causative agent, nBA, are presented. Broader implications for understanding the assembly of the RBC cytoskeleton-plasma membrane complexes, and their disassembly under normal, pathological, and forensic conditions are discussed.

We herein investigate the effects of varying the main experimental variables in one of the most used protocols for extracellular vesicle (EV) immobilisation on substrates for subsequent atomic force microscopy (AFM) quantitative morphometry and nanoindentation performed in liquid. We introduce the parameter Q as a quantitative measure of total adsorbed material and show how it can be used as an estimator of relative sample concentrations across multiple AFM imaging experiments. We show how Q is logarithmically dependent on substrate charge density, whereas the EV contact angle (CA) surprisingly does not follow the same dependence. Finally, we propose an optimised protocol for AFM quantitative morphometry in air that yields the same EV size distributions obtained in liquid.

Structured illumination microscopy (SIM) enables superresolution imaging of biological samples but suffers from artefacts, noise, and loss of high-frequency details in low-light conditions. These problems arise due to limitations in traditional reconstruction methods such as single-scale upsampling and pixel-wise losses that fail to capture SIM's multi-scale frequency patterns. We propose Att-SIM-LapSRN, a hybrid deep learning framework that integrates Attention U-Net with Laplacian pyramid super-resolution network (LapSRN) to address these challenges. Attention gates at skip connections selectively enhance salient feature representations corresponding to moiré patterns while attenuating background noise, producing sharper reconstructions precise localisation of cell structures. The LapSRN component employs progressive multiscale upsampling across pyramid levels to reduce the bicubic interpolation. Additionally, we introduce an FFT-based loss function that explicitly targets spatial frequency patterns, ensuring structural consistency, contrast enhancement and edge sharpness critical for SIM imaging. Our model was evaluated on the BioSR dataset, demonstrating superior performance over state-of-the-art methods, with significant improvements in PSNR, SSIM, and perceptual quality metrics. Att-SIM-LapSRN achieves enhanced lateral resolution and structural fidelity, making it a robust solution for high-quality SIM reconstruction in biological imaging applications.

Measuring the bismuth (Bi) content of ternary gallium arsenide bismuthide (GaAsBi) alloys is important because it sensitively influences their bandgap, and Bi is known to segregate vertically to the surface and sometimes also laterally during growth, so elemental distribution maps need to be quantified. A suitable method is mapping of characteristic X-rays based on energy-dispersive X-ray spectroscopy (EDXS) in a scanning transmission electron microscope (STEM). One of the key problems in this alloy system that there are several overlaps of characteristic X-ray lines from the corresponding elements, namely of As Kα with Bi Lα and of a sum peak of Ga Lα and As Lα with Bi Mα, which no standard solid-state detector could distinguish. Routine quantification procedures thus often fail, exhibiting unacceptably large scatter. Here, an iterative procedure using k*-factors is outlined, leading to improved quantification using sets of different X-ray line pairs to be consistent within better than 1% bismuth coverage of the group V sub-lattice, for a range up to 14%.