npj Imaging

The tissue-resident immune system involves complex 3D assemblies that interact with extended structures such as blood vessels and nerves. These interactions are difficult to study using conventional 2D profiling because they span many tissue sections. In animal tissues, volumetric imaging approaches such as light-sheet fluorescence microscopy (LSFM) are widely used to study 3D tissue organization, with labelling often aided by genetically encoded reporters and vascular dyes. In contrast, LSFM of human specimens remains underdeveloped because most clinical samples are available only as formalin-fixed paraffin-embedded (FFPE) tissue, limiting labeling strategies primarily to dyes and antibodies. Here, we present a volumetric cyclic immunofluorescence (v-CyCIF) and virtual H&E toolbox that overcomes key barriers to multiplexed imaging of immune cells and nerves in human specimens up to 1 mm thick. We use v-CyCIF to study neuroimmune interactions in normal and cancer tissues and to immunoprofile intact secondary and tertiary lymphoid structures. Re-embedding and sectioning of specimens following volumetric imaging enables high-plex high-resolution analysis of subcellular structures and cell-cell interactions associated with immune cell activity. v-CyCIF therefore provides a flexible framework for multi-scale 3D profiling of clinical specimens across imaging formats and resolutions.

Magnetic resonance imaging (MRI) is essential for visualizing the healthy and diseased brain, yet the cellular basis of MRI signal and how it changes over time remain poorly understood. Here, we present BRIDGE (Brain Radiological Imaging with Deep-learning based Ground-Truth Exploration), a platform integrating in vivo MRI with in vivo two-photon (2P) and ex vivo super-resolution microscopy using a multi-step, iterative co-registration pipeline. It enables in vivo, longitudinal, and voxel-precise mapping of MRI signals to their cellular origins for the first time. The registered overlay reveals the cellular and anatomical origins of MRI signals and enables training of convolutional neural networks to enhance the effective resolution of MRI. Using BRIDGE, we identified a microenvironmental vessel biomarker for early metastatic colonization in patient-derived xenograft models of brain metastasis. In particular we found that distinct T2*-weighted hypointense lesions correspond to reduced blood flow and erythrostasis in perimetastatic capillaries. In glioma, longitudinal intravital studies further demonstrated direct correlations between non-vasogenic T2-weighted signal changes and patient-dependent tumor growth dynamics. Taken together, BRIDGE advances radiological interpretation by establishing a microscopic ground truth for MRI signatures over time, enabling deep learning-based predictive histology, and providing cellular-level insights into tumor microenvironment features with direct clinical imaging implications. Graphical abstractBRIDGE enables longitudinal voxel-to-voxel correlation and ground truth based automatic segmentation of MR images O_FIG O_LINKSMALLFIG WIDTH=177 HEIGHT=200 SRC="FIGDIR/small/680637v1_ufig1.gif" ALT="Figure 1"> View larger version (73K): org.highwire.dtl.DTLVardef@f1f64eorg.highwire.dtl.DTLVardef@1619e3eorg.highwire.dtl.DTLVardef@1dc2e7forg.highwire.dtl.DTLVardef@7097e1_HPS_FORMAT_FIGEXP M_FIG C_FIG

Label-free optical redox imaging utilizes endogenous NAD(P)H and FAD autofluorescence to evaluate metabolism in living specimens. The conventional optical redox ratio collapses these two channels into a single value; however, it does not indicate whether a pixel has sufficient photon support or the cellular context necessary for quantitative aggregation. To address this limitation, we introduce FPhaS, a fixed-calibration phase- autofluorescence framework that integrates quantitative phase imaging (QPI) with simultaneous label-free autofluorescence multi-harmonic microscopy (SLAM), using fluorescence lifetime imaging (FLIM) solely for validation. Because QPI and SLAM are acquired with the same objective, a unified non-biological calibration aligns phase-derived structural data with the autofluorescence frame, yielding a residual error of 0.39 pixels. This calibration is maintained across all biological specimens. This shared geometric reference enables local evaluation of structural and metabolic information, rather than comparing approximately aligned images. FPhaS decomposes the data into cell presence, ratio credibility, and confidence-supported pooling. We validated FPhaS on A549 cells under high and low-photon conditions; the framework is designed to generalize to other cell and tissue types. Confidence-weighted intensity redox estimates were compared with lifetime-derived measurements within mask-locked cellular regions. Concordance improved exclusively when both the denominator photon support and an independent structural criterion were satisfied. The same reference layer generated cell-level descriptors of metabolic content, metabolic-structural organization, and measurement reliability, while also constraining the CombinedWLS reconstruction under diminished fluorescence acquisition. FPhaS redefines label-free metabolic imaging from producing comprehensive ratio maps to identifying regions where optical evidence substantiates quantitative inference.

The rapid convergence of optical innovation and machine intelligence is reshaping biological imaging by enabling platforms that jointly advance image formation and computational reconstruction for highspeed, high-resolution volumetric microscopy. However, broadly accessible three-dimensional imaging at high spatiotemporal resolution remains limited by the reliance of existing supervised methods on large modality-matched training datasets, the computational burden of conventional iterative reconstruction, and sensitivity to optical mismatch arising from small deviations in the spatial-angular point spread functions. Here, we introduce HYPER-Net, a physics-conditioned self-supervised framework for Fourier light-field microscopy that integrates scan-free volumetric acquisition with fast, robust three-dimensional reconstruction. HYPER-Net incorporates experiment-specific point-spread functions into the learning process in two complementary roles: as the forward operator that enforces measurement consistency and as a conditioning signal that adaptively modulates intermediate feature representations. This design reduces reliance on paired experimental ground-truth volumes, improves robustness to system variation, and enables generalizable reconstruction across diverse biological contexts. Using human colon organoids, embryonic Xenopus laevis hearts, hiPSC-derived cardiac spheroids, and freely moving Caenorhabditis elegans, we demonstrate high-fidelity volumetric imaging of tissue morphology, cardiac function, calcium-contraction coupling, and locomotion-associated neural and muscular dynamics. These results position HYPER-Net as a versatile framework for rapid volumetric imaging and quantitative analysis of dynamic biological systems across basic research and biomedical applications.

Deep learning provides an opportunity to automatically segment and extract cellular features from high-throughput microscopy images. Many labeling strategies have been developed for this purpose, ranging from the use of fluorescent markers to label-free approaches. However, differences in the channels available to each respective training dataset make it difficult to directly compare the effectiveness of these strategies across studies. Here we explore training models using subimage stacks composed of channels sampled from larger, hyper-labeled, image stacks. This allows us to directly compare a variety of labeling strategies and training approaches on identical cells. This approach revealed that fluorescence-based strategies generally provide higher segmentation accuracies but were less accurate than label-free models when labeling was inconsistent. The relative strengths of label and label-free techniques could be combined through the use of merging fluorescence channels and out-of-focus brightfield images. Beyond comparing labeling strategies, using subimage stacks for training was also found to provide a method of simulating a wide range of labeling conditions, increasing the ability of the final model to accommodate a greater range of experimental setups.

Cryo-electron tomography (CET) combined with sub-volume averaging (SVA), is currently the only imaging technique capable of determining the structure of proteins imaged inside cells at molecular resolution. To obtain high-resolution reconstructions, sub-volumes containing randomly distributed copies of the protein of interest need be identified, extracted and subjected to SVA, making accurate particle detection a critical step in the CET processing pipeline. Classical template-based methods have high false-positive rates due to the very low signal-to-noise ratios (SNR) typical of CET volumes, while more recent neural-network based detection algorithms require extensive labeling, are very slow to train and can take days to run. To address these issues, we propose a novel particle detection framework that uses positive-unlabeled learning and exploits the unique properties of 3D tomograms to improve detection performance. Our end-to-end framework is able to identify particles within minutes when trained using a single partially labeled tomogram. We conducted extensive validation experiments on two challenging CET datasets representing different experimental conditions, and observed more than 10% improvement in mAP and F1 scores compared to existing particle picking methods used in CET. Ultimately, the proposed framework will facilitate the structural analysis of challenging biomedical targets imaged within the native environment of cells.

In standard histopathology, hematoxylin and eosin (H&E) staining stands as a pivotal tool for cancer tissue analysis. However, this method is limited to two-dimensional (2D) analysis or requires labor-intensive preparation for three-dimensional (3D) inspection of cancer tissues. In this study, we present a method for 3D virtual H&E staining of label-free cancer tissues, employing holotomography and deep learning. Holotomography is used to measure the 3D refractive index (RI) distribution of the label-free cancer slides. A deep learning-based image-to-image translation framework is integrated into the resulting 3D RI distribution, enabling virtual H&E staining in 3D. Our method has been applied to colon cancer tissue slides with thicknesses up to 20 m, with conventional chemical H&E staining providing a direct validation for the method. This framework not only bypasses the conventional staining process but also provides 3D structures of glands, lumens, and individual nuclei. The results demonstrate enhancement in histopathological efficiency and the extension of the standard histopathology into the 3D realm. To validate the repeatability and scalability of the approach, we applied the framework to the gastric cancer slides obtained from different institute and imaging devices.

Coherent Raman spectroscopy enables label-free biochemical fingerprinting of live cells with subcellular resolution. We previously developed a machine learning framework capable of classifying glioma FFPE tissues using Raman spectral signatures. To accelerate live cell acquisition, we previously developed RADAR (Raman Spectral Analysis Using Deep Learning for Artifact Removal), a method that increases imaging speed by an order of magnitude while preserving spectral integrity. By integrating high-speed Raman imaging with supervised machine learning, we aimed to define unique biochemical fingerprints specific to cell type. We hypothesized that intrinsic biochemical composition alone is sufficient to distinguish cellular identity and tumor subtype. To test this, we generated metabolic maps of diverse brain-derived cell types--including astrocytoma, oligodendroglioma, and glioblastoma cells--using coherent Raman spectroscopy at single-cell resolution. Patient-derived brain tumor cell lines representing genetically heterogeneous backgrounds were analyzed. Samples were stratified by IDH1 mutation status (IDH1-mutant and IDH1-wild-type) and histologically classified as oligodendroglioma or astrocytoma. Raman spectral data were acquired from 286 live single cells across the two principal molecular classes, with further subdivision into two histologic subtypes within the IDH1-mutant group. Classification was performed using an XGBoost model with shallow tree depth (1-3), a 20% held-out test set, and grouped, stratified 5-fold cross-validation to control for sample-level bias. The machine learning framework distinguished IDH1-mutant from IDH1-wild-type cells with a ROC-AUC of 0.78 and further discriminated IDH1-mutant astrocytoma from oligodendroglioma cells with a ROC-AUC of 0.81. Feature importance analysis demonstrated that separation between IDH1-mutant and IDH1-wild-type cells was driven primarily by Raman peaks associated with protein amide bands, total NADH, unsaturated fatty acids, and heme-related vibrational modes. Within the IDH1-mutant class, discrimination between oligodendroglioma and astrocytoma was driven by lipid-rich vesicle signatures, protein/polyamide amide bands, and lipid-associated spectral features. Together, these findings support the feasibility of label-free, machine learning-assisted Raman profiling to resolve clinically relevant glioma subtypes at single-cell resolution. This scalable analytical framework provides a translational platform for investigating metabolic heterogeneity, therapeutic response, co-culture systems, and patient-derived organoid models.

Three-dimensional fluorescence microscopy often exhibits anisotropic resolution because axial information is poorly sampled and more blurred than lateral information, which complicates quantitative interpretation of fine 3D structures. Although optical remedies and computational restoration have been explored, many approaches require demanding system calibration or rely on accurate PSF models and assumptions that are difficult to satisfy across all samples and modalities. Here we present DeepIso, a self-supervised isotropy restoration framework that couples supervised pretraining with an internal-learning inference stage to estimate degradation directly from the measured volume. Without explicit PSF specification or enforced lateral-axial structural equivalence, DeepIso recovers axial frequency content and improves the continuity of elongated structures while retaining fine features, with superior performance over existing computational approaches in terms of both visual inspection and quantitative metrics. The method is validated on synthetic benchmarks and experimental datasets, demonstrating isotropy enhancement across confocal, light-sheet, and 3D structured illumination microscopy, thereby supporting downstream volumetric analysis including segmentation and tracking.

Mitochondria, acting as the energy powerhouse, biosynthetic center, and reductive equivalent hub of the cell, participate in cellular metabolic activities. However directly imaging mitochondrial chemical content and quantifying metabolic activity in living cells remain challenging. Here, by Self-PErmutation Noise2noise Denoiser enhanced Hyperspectral Stimulated Raman Scattering (SPEND-hSRS) microscopy, we demonstrate fingerprint-region metabolic imaging of fumarate, a key intermediate in the tricarboxylic acid (TCA) cycle, with sub-millimolar sensitivity. In chemotherapy-stressed bladder cancer cells, fumarate imaging revealed two mitochondrial subpopulations with divergent TCA metabolic preferences quantified by ratio metric analysis. Pixel-wise least absolute shrinkage and selection operator (LASSO) spectral unmixing further reconstructs fumarate and lipid maps, uncovering localized fumarate enrichment in protrusions. Extending to CH-window hyperspectral SRS imaging, we uncover the interplay between mitochondria and lipid droplets (LDs) in protrusions, where fatty acid is found to be released from LDs, to fuel the TCA cycle. Together, our work establishes SPEND-hSRS as high-resolution platform for linking fumarate to mitochondrial heterogeneity. Our results provide new insights into how mitochondrial heterogeneity and interaction with LDs drive cancer cell adaptation to stress.

The resolving power of optical microscopy is fundamentally constrained by the diffraction of light, limiting our ability to visualize subcellular structures. Computational methods, particularly deconvolution, can restore blurred images but critically depend on an accurate point spread function (PSF), whose estimation is often impractical and error-prone, leading to artifacts. Here, we introduce Nonlinear Fourier Re-weighting (NFR), a rapid algorithm that operates without any prior knowledge of the imaging system, achieving deconvolution-like effects through a single logarithmic mapping of the images Fourier spectrum. This non-iterative process re-balances spatial frequency components to computationally reverse the effects of optical blurring. We demonstrate that NFR robustly enhances resolution beyond the Sparrow limit and recovers authentic structural details. NFR excels where traditional methods fail, remaining effective in the presence of severe optical aberrations and high noise. Furthermore, NFR synergistically improves the output of super-resolution modalities like structured illumination microscopy (SIM), and its near-instantaneous processing enables real-time enhancement of dynamic biological processes, such as in vivo multi-photon microscopic imaging deep within scattering tissue. By decoupling high-fidelity image restoration from system modeling, NFR offers a powerful, accessible, and universally applicable tool for improving image quality across diverse microscopic techniques, facilitating the analysis of large datasets and the discovery of previously obscured biological phenomena.

Advancements in transmission electron microscopy (TEM) have enabled in-depth studies of biological specimens, offering new avenues to large-scale imaging experiments with subcellular resolution. Mitochondrial structure is of growing interest in cancer biology due to its crucial role in regulating the multi-faceted functions of mitochondria. We and others have established the crucial role of mitochondria in triple-negative breast cancer (TNBC), an aggressive subtype of breast cancer with limited therapeutic options. Building upon our previous work demonstrating the regulatory role of mitochondrial structure dynamics in the metabolic adaptations and survival of chemotherapy-refractory TNBC cells, we sought to extend those findings to a large-scale analysis of transmission electron micrographs. Here we present a U-Net artificial intelligence (AI) model for automatic annotation and assessment of mitochondrial morphology and feature quantification. Our model is trained on 11,039 manually annotated mitochondria across 125 micrographs derived from a variety of orthotopic patient-derived xenograft (PDX) mouse model tumors and adherent cell cultures. The model achieves an F1 score of 0.85 on test micrographs at the pixel level. To validate the ability of our model to detect expected mitochondrial structural changes, we utilized micrographs from mouse primary skeletal muscle cells genetically modified to lack Dynamin-related protein 1 (Drp1). The algorithm successfully detected a significant increase in mitochondrial elongation, in alignment with the well-established role of Drp1 as a driver of mitochondrial fission. Further, we subjected in vitro and in vivo TNBC models to conventional chemotherapy treatments commonly used for clinical management of TNBC, including doxorubicin, carboplatin, paclitaxel, and docetaxel (DTX). We found substantial within-sample heterogeneity of mitochondrial structure in both in vitro and in vivo TNBC models and observed a consistent reduction in mitochondrial elongation in DTX-treated specimens. We went on to compare mammary tumors and matched lung metastases in a highly metastatic PDX model of TNBC, uncovering significant increase in mitochondrial length in metastatic lesions compared to their cognate mammary tumor. This dataset provides high statistical power to detect frequent chemotherapy-induced shifts in mitochondrial shapes and sizes in residual cells left behind after treatment. The successful application of our AI model to capture mitochondrial structure marks a step forward in high-throughput analysis of mitochondrial structures, enhancing our understanding of how morphological changes may relate to chemotherapy efficacy and mechanism of action. Our large, manually curated electron micrograph dataset - now publicly available - serves as a unique gold-standard resource for developing, benchmarking, and applying computational models, while further advancing investigations into mitochondrial morphology and its impact on breast cancer biology.

Light-sheet fluorescence microscopy (LSFM) has revolutionized biological imaging by enabling high spatial and temporal resolution with minimal photodamage. However, conventional LSFM techniques often suffer from striping artifacts in the resulting images due to light scattering and absorption within samples, leading to uneven illumination that negatively impacts the accuracy of subsequent image analyses. To address this limitation, we introduce dodecagon light-sheet fluorescence microscopy (dodecaLSFM), a novel approach that maximizes angular diversity to achieve homogeneous illumination and suppress striping artifacts. dodecaLSFM employs diffraction optics and cylindrical lenses to generate twelve light sheets, providing 360 degree omnidirectional illumination that significantly enhances illumination uniformity compared to traditional mSPIM, mDSLM, and ultramicroscopy systems, which use only one or two illumination planes. We demonstrate the effectiveness of dodecaLSFM by achieving high-resolution imaging of whole mouse brain vasculature following tissue clearing, allowing precise morphometric analysis of vascular networks without striping artifacts. Furthermore, we show that combining dodecaLSFM with expansion microscopy (ExM) enables whole-organ 3D imaging at cellular resolution. This novel approach provides an advanced, scalable solution for large-volume imaging, facilitating detailed structural and functional studies across diverse biological applications.

Quantitative molecular imaging in photoacoustics is fundamentally limited by the ill-posed nature of spectral unmixing, where spectral overlap, noise, and unknown fluence introduce bias in conventional inversion-based methods. We introduce photoacoustic fingerprinting (PAF), a framework that reframes spectral unmixing as a fingerprint recognition problem. PAF interprets multispectral signals as high-dimensional fingerprints encoding both molecular composition and measurement distortions. Inspired by magnetic resonance fingerprinting, PAF uses a recurrent neural network trained on synthetic data spanning realistic mixtures, noise levels, and fluence variations to directly infer molecular concentrations from spectral shape. PAF enables accurate and robust quantification in regimes where conventional methods break down, including low signal-to-noise conditions, spectrally correlated mixtures, and unknown fluence distortions. In controlled simulations, PAF consistently outperformed non-negative least squares, with the largest gains observed for spectrally overlapping chromophores such as collagen. In phantom studies, PAF improved molecular specificity by correctly localizing collagen and recovering water contrast despite similar spectral reconstructions. In ex vivo mouse livers, PAF detected lipid accumulation associated with steatosis, and in human arteries, it identified molecular signatures consistent with thrombus and lipid-rich plaque. These results establish PAF as a generalizable framework for label-free molecular imaging and a promising step toward quantitative photoacoustic diagnostics.

Investigating cardiac dynamics, including contractile function and intracardiac flow, requires volumetric imaging capable of resolving whole-organ events at micrometer resolution and millisecond timescales. However, the limited readout bandwidth of detectors imposes fundamental trade-offs among spatial sampling, field of view, and achievable volume rates. Here we introduce compressive axial-integrated planar scanning (CAPS) microscopy, a computational imaging framework that combines rapid light-sheet scanning, detection-side axial multiplexing with model-based reconstruction to enhance detector bandwidth utilization for high-speed volumetric imaging. Using widely accessible optical sensors and components, CAPS achieves cellular-scale resolving power across heart chambers at 200 volumes per second with an effective detector pixel rate of 5.82 GHz, representing a [~]15-fold increase in spatiotemporal throughput relative to uncompressed volumetric acquisition. Coordinated high-speed encoding and computational reconstruction further mitigate rolling-shutter distortions in CMOS sensors while preserving frame rate and intrinsic optical sectioning. We demonstrate that CAPS enables beat-resolved imaging of single-cell cardiomyocyte kinematics, chamber-scale contractile dynamics, and intracardiac hemodynamics in zebrafish larvae under both healthy and pharmacologically perturbed conditions. Collectively, these advances establish CAPS as a powerful framework for quantitative, in vivo characterization of coordinated and disrupted cardiac dynamics at cellular resolution, supporting high-speed volumetric interrogation of organ-level function and disease progression.

Optical fluorescence microscopy enables visualization of biological structures and dynamics. However, the intrinsic diffraction limit, especially axially, and depth-related scattering noise compromise the image resolution and fidelity. Computational 3D deconvolution is a promising approach for mitigating these issues, yet its execution is hindered by inaccurate and cumbersome theoretical modeling or experimental measurement of 3D point spread function (PSF), as well as ineffective 3D noise regularization. Furthermore, in the 3D super-resolution regime, there remains a lack of standardized tools for evaluating 3D super-resolution fidelity. Here, we present the 3D adaptive deconvolution and evaluation (3D-ADE) toolkit, which comprises 3D-Ada deconvolution with physics-oriented automatic 3D-PSF calibration, and 3D-SQUIRREL for 3D super-resolution quality assessment. It effectively resolves noise instability, eliminates the need for 3D-PSF calibration, and reliably assesses the fidelity of 3D resolution extension via deconvolution, physical, and deep-learning-based methods. Accessible via multiple software platforms, 3D-ADE enhances the versatility of 3D deconvolution and fills the gap in 3D super-resolution evaluation tools, and thereby advances volumetric fluorescence imaging applications.

Artificial vision systems hold transformative potential for biomedical imaging, diagnostics, and translational research by emulating and extending the capabilities of biological eyes. However, current techniques often face intrinsic trade-offs between spatial resolution, field of view, and depth perception, particularly in compact, biologically relevant settings. Here, we introduce FOLIC, a foveated light-field compound imaging system, which integrates compound-eye-inspired wide angular coverage and chambered-eye-inspired spatial acuity within a unified multi-aperture concave architecture. FOLIC naturally generates peripheral, blend, and foveated zones from a single capture, enabling seamless, depth-extended, multiscale visualization from wide-field context down to single-cell lateral resolution. We validate FOLIC across diverse fluorescent and non-fluorescent specimens, including cellular phantoms, tissue sections, and small organisms, demonstrating its versatility and scalability for biomedical research and related translational applications. We anticipate FOLIC to offer a biologically informed design blueprint for future artificial vision systems. TeaserA bioinspired system unifies compound and chambered eye principles to achieve wide-field volumetric microscopy.

Single-Molecule Localization Microscopy remains limited in its ability for robust and simple multi-color imaging. Whereas the fluorophore Alexa647 is widely used due to its brightness and excellent blinking dynamics, other excellent blinking fluorophores, such as CF660 and CF680, spectrally overlap. Here we present Probability-based Fluorophore Classification, a method to perform multi-color SMLM with Alexa647, CF660 and CF680 that uses statistical decision theory for optimal classification. The emission is split in a short and long wavelength channel to enable classification and localization without a major loss in localization precision. Each emitter is classified using a Generalized Maximum Likelihood Ratio Test using the photon statistics of both channels. This easy-to-adopt approach does not require nanometer channel registration, is able to classify fluorophores with tunable low false positive rates (<0.5%) and optimal true positive rates and outperforms traditional ratiometric spectral de-mixing and Salvaged Fluorescence. We demonstrate its applicability on a variety of samples and targets.

Advances in multi-dimensional imaging method and probe developments have brought super-resolution fluorescence microscopy into a functional era. They capture additional single-molecule fluorescence information concurrently with spatial localization, enabling simultaneous identification of molecular species and interrogation of nanoscale environments with rich, high-dimensional imaging information. However, the adoption of multi-dimensional imaging has been hindered by fragmented analysis workflows, complex parameter tuning, and limited integration of advanced computational methods. Here, we introduce an agentic single-molecule multi-dimensional bioimaging AI, referred to as SIMBA, an AI-driven platform that unifies single-molecule localization, spectral processing and deep learning-based denoising within a single agentic and interactive framework. SIMBA incorporates large language model-based agents capable of interpreting user intent, orchestrating analysis pipelines, and dynamically selecting computational tools for automated data processing. We demonstrate that SIMBA enables supports standard single-molecule localization workflow, functional mapping of nanoscale environmental heterogeneity through single-molecule spectral analysis and denoising using developed supervised learning methods. By integrating extensible tool architectures with human language-guided workflows, SIMBA establishes a new paradigm for intelligent microscopy analysis, lowering barriers to multi-dimensional imaging adoption while enabling scalable, reproducible, and adaptive analysis of complex imaging datasets.

Localization microscopy techniques have overcome the diffraction limit, enabling nanoscale biological imaging by precisely determining the positions of individual emitters. However, the performance of deep learning methods commonly applied to these tasks often depends significantly on the quality of training data, typically generated through simulation. Creating simulations that perfectly replicate experimental conditions remains challenging, resulting in a persistent simulation-to-experiment (sim2exp) gap. To bridge this gap, we propose a physics-informed generative model leveraging self-supervised learning directly on experimental data. Our model extends the Deep Latent Particles (DLP) framework by incorporating a physical Point Spread Function (PSF) model into the decoder, enabling it to disentangle learned realistic environments from precise emitter properties. Trained directly on unlabeled experimental images, our model intrinsically captures realistic background, noise patterns, and emitter characteristics. The decoder thus acts as a high-fidelity generator, producing fully labeled, realistic training images with known emitter locations. Using these generated datasets significantly improves the performance of supervised localization algorithms, particularly in challenging scenarios such as complex backgrounds and low signal-to-noise ratios. Our results demonstrate substantial improvements in localization accuracy and emitter detection, underscoring the practical benefit of our approach for real-world microscopy applications. We will make our code publicly available.