Large field-of-view non-invasive imaging through scattering layers using fluctuating random illumination

Fluorescence microscopy through scattering media with robust matrix factorization

Biological tissues, as natural scattering media, inherently disrupt structural information, presenting significant challenges for optical imaging. Complex light propagation through tissue severely degrades image quality, limiting conventional fluorescence imaging techniques to superficial depths. Extracting meaningful information from random speckle patterns is, therefore, critical for deeper tissue imaging. In this study, we present RNP (robust non-negative principal matrix factorization), an approach that enables fluorescence microscopy under diverse scattering conditions. By integrating robust feature extraction with non-negativity constraints, RNP effectively addresses challenges posed by non-sparse signals and background interference in scattering tissue environments. The framework operates on a standard epi-fluorescence platform, eliminating the need for complex instrumentation or precise alignment. The results from imaging scattered cells and tissues demonstrate substantial improvements in robustness, field of view, depth of field, and image clarity. We anticipate that RNP will become a valuable tool for overcoming scattering challenges in fluorescence microscopy and driving advancements in biomedical research.

High-fidelity single-pixel imaging through scattering media using quantum-state encoded illumination

Imaging through scattering media remains a challenge due to the loss of light intensity and phase, leading to poorly reconstructed images with high noise and low resolution. Here, we demonstrate a new approach, to the best of our knowledge, to image through scattering media with unprecedented grayscale accuracy by encoding quantum states within illumination patterns. The orthogonality of quantum states significantly reduces cross-pixel interference, increases grayscale accuracy, and simplifies image reconstruction. We achieve a spatial density of 180 dots per inch with excellent grayscale linearity. Our method employs a single photodetector and a consumer-grade LED projector, enabling low-cost, noninvasive imaging through scattering media.

Noninvasive megapixel fluorescence microscopy through scattering layers by a virtual incoherent reflection matrix

Optical-resolution fluorescence imaging through and within complex samples presents a major challenge due to random light scattering, with substantial implications across multiple fields. While considerable advancements in coherent imaging through severe multiple scattering have been recently introduced by reflection matrix processing, approaches that tackle scattering in incoherent fluorescence imaging have been limited to sparse targets, require high-resolution control of the illumination or detection wavefronts, or require a very large number of measurements. Here, we present an approach that allows the adaptation of well-established reflection matrix techniques to scattering compensation in incoherent fluorescence imaging. We experimentally demonstrate that a small number of conventional wide-field fluorescence microscope images acquired under unknown random illuminations can effectively be used to construct a virtual fluorescence-based reflection matrix. Processing this matrix by an adapted matrix-based scattering compensation algorithm allows reconstructing megapixel-scale images from <150 acquired frames, without any spatial light modulators or computationally intensive processing.

Whispering Gallery Mode Micro/Nanolasers for Intracellular Probing at Single Cell Resolution

Intracellular probing at single cell resolution is key to revealing the heterogeneity of cells, learning new cell subtypes and functions, understanding the pathophysiology of disease, and ensuring precise diagnosis and treatment. Despite the best efforts, an enormous challenge remains due to the very small size, extremely low content, and dynamic microenvironment of a single cell. Whispering gallery mode (WGM) micro/nanolasers (active WGM) offer unique advantages of small mode volume, high quality factors, bright and low threshold laser emission, and narrow line width, particularly suitable for integration within a single cell. In this review, we provide a focused overview of WGM micro/nanolasers for intracellular probing. We deliver information on WGM micro/nanolaser concepts, sensing mechanism, and biocompatibility, as well as recent progress in intracellular probing applications mainly covering cellular-level sensing, molecular-level detection, and feasibility for cellular imaging. At the end, challenges and prospects of WGM micro/nanolasers for intracellular applications are discussed.

Large-scale scattering-augmented optical encryption

Data proliferation in the digital age necessitates robust encryption techniques to protect information privacy. Optical encryption leverages the multiple degrees of freedom inherent in light waves to encode information with parallel processing and enhanced security features. However, implementations of large-scale, high-security optical encryption have largely remained theoretical or limited to digital simulations due to hardware constraints, signal-to-noise ratio challenges, and precision fabrication of encoding elements. Here, we present an optical encryption platform utilizing scattering multiplexing ptychography, simultaneously enhancing security and throughput. Unlike optical encoders which rely on computer-generated randomness, our approach leverages the inherent complexity of light scattering as a natural unclonable function. This enables multi-dimensional encoding with superior randomness. Furthermore, the ptychographic configuration expands encryption throughput beyond hardware limitations through spatial multiplexing of different scatterer regions. We propose a hybrid decryption algorithm integrating model- and data-driven strategies, ensuring robust decryption against various sources of measurement noise and communication interference. We achieved optical encryption at a scale of ten-megapixel pixels with 1.23 µm resolution. Communication experiments validate the resilience of our decryption algorithm, yielding high-fidelity results even under extreme transmission conditions characterized by a 20% bit error rate. Our encryption platform offers a holistic solution for large-scale, high-security, and cost-effective cryptography.

Neurophotonics beyond the surface: unmasking the brain's complexity exploiting optical scattering

The intricate nature of the brain necessitates the application of advanced probing techniques to comprehensively study and understand its working mechanisms. Neurophotonics offers minimally invasive methods to probe the brain using optics at cellular and even molecular levels. However, multiple challenges persist, especially concerning imaging depth, field of view, speed, and biocompatibility. A major hindrance to solving these challenges in optics is the scattering nature of the brain. This perspective highlights the potential of complex media optics, a specialized area of study focused on light propagation in materials with intricate heterogeneous optical properties, in advancing and improving neuronal readouts for structural imaging and optical recordings of neuronal activity. Key strategies include wavefront shaping techniques and computational imaging and sensing techniques that exploit scattering properties for enhanced performance. We discuss the potential merger of the two fields as well as potential challenges and perspectives toward longer term in vivo applications.

Roadmap on computational methods in optical imaging and holography [invited]

Computational methods have been established as cornerstones in optical imaging and holography in recent years. Every year, the dependence of optical imaging and holography on computational methods is increasing significantly to the extent that optical methods and components are being completely and efficiently replaced with computational methods at low cost. This roadmap reviews the current scenario in four major areas namely incoherent digital holography, quantitative phase imaging, imaging through scattering layers, and super-resolution imaging. In addition to registering the perspectives of the modern-day architects of the above research areas, the roadmap also reports some of the latest studies on the topic. Computational codes and pseudocodes are presented for computational methods in a plug-and-play fashion for readers to not only read and understand but also practice the latest algorithms with their data. We believe that this roadmap will be a valuable tool for analyzing the current trends in computational methods to predict and prepare the future of computational methods in optical imaging and holography.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00340-024-08280-3.

Memory-less scattering imaging with ultrafast convolutional optical neural networks

The optical memory effect in complex scattering media including turbid tissue and speckle layers has been a critical foundation for macroscopic and microscopic imaging methods. However, image reconstruction from strong scattering media without the optical memory effect has not been achieved. Here, we demonstrate image reconstruction through scattering layers where no optical memory effect exists, by developing a multistage convolutional optical neural network (ONN) integrated with multiple parallel kernels operating at the speed of light. Training this Fourier optics-based, parallel, one-step convolutional ONN with the strong scattering process for direct feature extraction, we achieve memory-less image reconstruction with a field of view enlarged by a factor up to 271. This device is dynamically reconfigurable for ultrafast multitask image reconstruction with a computational power of 1.57 peta-operations per second (POPS). Our achievement establishes an ultrafast and high energy-efficient optical machine learning platform for graphic processing.

Multimodal 2D and 3D microscopic mapping of growth cartilage by computational imaging techniques - a short review including new research

Being able to image the microstructure of growth cartilage is important for understanding the onset and progression of diseases such as osteochondrosis and osteoarthritis, as well as for developing new treatments and implants. Studies of cartilage using conventional optical brightfield microscopy rely heavily on histological staining, where the added chemicals provide tissue-specific colours. Other microscopy contrast mechanisms include polarization, phase- and scattering contrast, enabling non-stained or 'label-free' imaging that significantly simplifies the sample preparation, thereby also reducing the risk of artefacts. Traditional high-performance microscopes tend to be both bulky and expensive.Computational imagingdenotes a range of techniques where computers with dedicated algorithms are used as an integral part of the image formation process. Computational imaging offers many advantages like 3D measurements, aberration correction and quantitative phase contrast, often combined with comparably cheap and compact hardware. X-ray microscopy is also progressing rapidly, in certain ways trailing the development of optical microscopy. In this study, we first briefly review the structures of growth cartilage and relevant microscopy characterization techniques, with an emphasis on Fourier ptychographic microscopy (FPM) and advanced x-ray microscopies. We next demonstrate with our own results computational imaging through FPM and compare the images with hematoxylin eosin and saffron (HES)-stained histology. Zernike phase contrast, and the nonlinear optical microscopy techniques of second harmonic generation (SHG) and two-photon excitation fluorescence (TPEF) are explored. Furthermore, X-ray attenuation-, phase- and diffraction-contrast computed tomography (CT) images of the very same sample are presented for comparisons. Future perspectives on the links to artificial intelligence, dynamic studies andin vivopossibilities conclude the article.

Non-invasive accelerated imaging through a scattering medium via multi-stage complexity guidance

The research of scattering imaging is of great significance to the development of various fields, but the existing scattering imaging methods are difficult to combine for the advantages of non-invasiveness, real-time imaging, and high quality. In this paper, a new, to our knowledge, scattering imaging technique is proposed that optimizes the traditional autocorrelation imaging technique by multi-stage complexity guidance and the initial acceleration module. We introduce the complexity difference index into the phase iterative recovery step for effective complexity guidance, and add the initial module based on error-reduction iteration to realize a fast startup. A series of experiments is carried out to test the performance of the new technique. The results show that the proposed technique significantly improves the scattering reconstruction speed. Meanwhile, the accuracy and clarity of the reconstructed image are significantly higher than the traditional method in terms of fast imaging. Moreover, this technique has better robustness to noise compared to the traditional autocorrelation imaging technique. The experimental code for this paper is available on GitHub.

Widefield functional speckle-correlation optical scattering mesoscopy toward hemodynamic imaging

Speckle-correlation optical scattering imaging (SCOSI) has shown the potential for non-invasive biomedical diagnostic applications, which directly utilizes the scattering patterns to reconstruct the deep and non-line-of-sight objects. However, the course of the translation of this technique to preclinical biomedical imaging applications has been postponed by the following two facts: 1) the field of view of SCOSI was significantly limited by the optical memory effect, and 2) the molecular-tagged functional imaging of the biological tissues remains largely unexplored. In this work, a proof-of-concept design of the first-generation widefield functional SCOSI (WF-SCOSI) system was presented for simultaneously achieving mesoscopic mapping of fluid morphology and flow rate, which was realized by implementing the concepts of scanning synthesis and fluorescence scattering flowmetry. The ex vivo imaging results of the fluorescence-labeled large-scale blood vessel network phantom underneath the strong scatters demonstrated the effectiveness of WF-SCOSI toward non-invasive hemodynamic imaging applications.

Neurophotonics beyond the Surface: Unmasking the Brain's Complexity Exploiting Optical Scattering

The intricate nature of the brain necessitates the application of advanced probing techniques to comprehensively study and understand its working mechanisms. Neurophotonics offers minimally invasive methods to probe the brain using optics at cellular and even molecular levels. However, multiple challenges persist, especially concerning imaging depth, field of view, speed, and biocompatibility. A major hindrance to solving these challenges in optics is the scattering nature of the brain. This perspective highlights the potential of complex media optics, a specialized area of study focused on light propagation in materials with intricate heterogeneous optical properties, in advancing and improving neuronal readouts for structural imaging and optical recordings of neuronal activity. Key strategies include wavefront shaping techniques and computational imaging and sensing techniques that exploit scattering properties for enhanced performance. We discuss the potential merger of the two fields as well as potential challenges and perspectives toward longer term in vivo applications.

3D fluorescence imaging through scattering medium using transport of intensity equation and iterative phase retrieval

In this paper, we have proposed a method of three-dimensional (3D) fluorescence imaging through a scattering medium. The proposed method combines the numerical digital phase conjugation propagation after measurement of the complex amplitude distribution of scattered light waves by the transport of intensity equation (TIE) with followed iterative phase retrieval to achieve 3D fluorescence imaging through a scattering medium. In the experiment, we present the quantitative evaluation of the depth position of fluorescent beads. In addition, for time-lapse measurement, cell division of tobacco-cultured cells was observed. Numerical results presented the effective range of the phase amount in the scattering medium. From these results, the proposed method is capable of recovering images degraded by a thin scattering phase object beyond a small phase change approximation.

Single-pixel transmission matrix recovery via two-photon fluorescence

Imaging at depth in opaque materials has long been a challenge. Recently, wavefront shaping has enabled notable advance for deep imaging. Nevertheless, most noninvasive wavefront-shaping methods require cameras, lack the sensitivity for deep imaging under weak optical signals, or can only focus on a single "guidestar." Here, we retrieve the transmission matrix (TM) noninvasively using two-photon fluorescence exploiting a single-pixel detection combined with a computational framework, allowing to achieve single-target focus on multiple guidestars spread beyond the memory effect range. In addition, if we assume that memory effect correlations exist in the TM, we are able to substantially reduce the number of measurements needed.

Super-resolution imaging through scattering media based on improved triple correlation recursion and deterministic iterative estimation

Iterative phase retrieval algorithms are commonly used in computational techniques and optimization methods to obtain the reconstruction of objects hidden behind opaque scattering media. However, these methods are susceptible to converging to incorrect local minima, and the calculation results tend to be unstable. In this paper, a triple-correlation-based super-resolution imaging (TCSI) framework is proposed to achieve single-shot imaging of unknown objects hidden behind the scattering medium. The amplitude spectrum of the object is obtained by a speckle correlation (SC) method. Iterative relaxation recursion (IRR) sufficiently extracts object information from the triple correlation (TC) of the speckle patterns, serving as the prior initial guess for the iterative estimation algorithm (IE) to obtain a deterministic phase spectrum. Blur correction (BC) is then applied to the diffraction-limited image to achieve super-resolution imaging. Experimental results demonstrate that the flexible framework could effectively overcome the influence of speckle resolution and outperform traditional methods in terms of performance. Our approach provides a basis for non-invasively visualizing various samples behind scattering media.

Learning diffractive optical communication around arbitrary opaque occlusions

Free-space optical communication becomes challenging when an occlusion blocks the light path. Here, we demonstrate a direct communication scheme, passing optical information around a fully opaque, arbitrarily shaped occlusion that partially or entirely occludes the transmitter's field-of-view. In this scheme, an electronic neural network encoder and a passive, all-optical diffractive network-based decoder are jointly trained using deep learning to transfer the optical information of interest around the opaque occlusion of an arbitrary shape. Following its training, the encoder-decoder pair can communicate any arbitrary optical information around opaque occlusions, where the information decoding occurs at the speed of light propagation through passive light-matter interactions, with resilience against various unknown changes in the occlusion shape and size. We also validate this framework experimentally in the terahertz spectrum using a 3D-printed diffractive decoder. Scalable for operation in any wavelength regime, this scheme could be particularly useful in emerging high data-rate free-space communication systems.

Imaging through a square multimode fiber by scanning focused spots with the memory effect

The existence of a shift-shift memory effect in square waveguides, whereby any translation of the input field induces translations in the output field in four symmetrical directions, has been previously observed by correlation measurements. Here we demonstrate that this memory effect is also observed in real space and can be put to use for imaging purposes. First, a focus is created at the output of a square-core multimode fiber, by wavefront shaping based on feedback from a guide-star. Then, because of the memory effect, four symmetrical spots can be scanned at the fiber output by shifting the wavefront at the fiber input. We demonstrate that this property can be exploited to perform fluorescence imaging through the multimode fiber, without requiring the measurement of a transmission matrix.

Characterization of the angular memory effect of dynamic turbid media

The optical angular memory effect (AME) is a basic feature of turbid media and defines the correlation of speckles when the incident light is tilted. AME based imaging through solid scattering media such as ground glass and biomedical tissue has been recently developed. However, in the case of liquid media such as turbid water or blood, the speckle pattern exhibits dynamic time-varying characteristics, which introduces several challenges. The AME of the thick volume dynamic media is particularly different from the layer scatterers. In practice, there are more parameters, e.g., scattering particle size, shape, density, or even the illuminating beam aperture that can influence the AME range. Experimental demonstration of AME phenomenon in liquid dynamic media and confirm the distinctions will contribution to complete the AME theory. In this paper, a dual-polarization speckle detection setup was developed to characterize the AME of dynamic turbid media, where two orthogonal polarized beams were employed for simultaneous detection by a single CCD. The AME of turbid water, milk and blood were measured. The influence of thickness, concentration, particle size and shape, and beam diameter were analyzed. The AME increasement of upon the decrease of beam diameter was tested and verified. The results demonstrate the feasibility of this method for investigating the AME phenomenon and provide guidance for AME based imaging through scattering media.

Complex amplitude field recovery of a scattering media obstructed object with multi-captured images

An iterative-based method for recovering the complex amplitude field behind scattering media is presented in this Letter. This method compensates the random phase modulation of scattering media by using multiple captured scattered light fields. Complex amplitude reconstruction with local iterative averaging of scattered light fields, and double weighted feedback is efficiently applied. Two feasible types of system setups, with varying detector positions and wavelength, are proposed. Simulations and proof-of-concept experiments are employed to demonstrate the effectiveness of the proposed method in reconstructing complex amplitude of a hidden target.

Speckle-correlation imaging through a kaleidoscopic multimode fiber

Speckle-correlation imaging techniques are widely used for noninvasive imaging through complex scattering media. While light propagation through multimode fibers and scattering media share many analogies, reconstructing images through multimode fibers from speckle correlations remains an unsolved challenge. Here, we exploit a kaleidoscopic memory effect emerging in square-core multimode fibers and demonstrate fluorescence imaging with no prior knowledge on the fiber. Experimentally, our approach simply requires to translate random speckle patterns at the input of a square-core fiber and to measure the resulting fluorescence intensity with a bucket detector. The image of the fluorescent object is then reconstructed from the autocorrelation of the measured signal by solving an inverse problem. This strategy does not require the knowledge of the fragile deterministic relation between input and output fields, which makes it promising for the development of flexible minimally invasive endoscopes.

Functional imaging through scattering medium via fluorescence speckle demixing and localization

Recently, fluorescence-based optical techniques have emerged as a powerful tool to probe information in the mammalian brain. However, tissue heterogeneities prevent clear imaging of deep neuron bodies due to light scattering. While several up-to-date approaches based on ballistic light allow to retrieve information at shallow depths inside the brain, non-invasive localization and functional imaging at depth still remains a challenge. It was recently shown that functional signals from time-varying fluorescent emitters located behind scattering samples could be retrieved by using a matrix factorization algorithm. Here we show that the seemingly information-less, low-contrast fluorescent speckle patterns recovered by the algorithm can be used to locate each individual emitter, even in the presence of background fluorescence. We test our approach by imaging the temporal activity of large groups of fluorescent sources behind different scattering phantoms mimicking biological tissues, and through a brain slice with a thickness of ∼200 µm.

Decision of the Optimal Rank of a Nonnegative Matrix Factorization Model for Gene Expression Data Sets Utilizing the Unit Invariant Knee Method: Development and Evaluation of the Elbow Method for Rank Selection

BACKGROUND

There is a great need to develop a computational approach to analyze and exploit the information contained in gene expression data. The recent utilization of nonnegative matrix factorization (NMF) in computational biology has demonstrated the capability to derive essential details from a high amount of data in particular gene expression microarrays. A common problem in NMF is finding the proper number rank (r) of factors of the degraded demonstration, but no agreement exists on which technique is most appropriate to utilize for this purpose. Thus, various techniques have been suggested to select the optimal value of rank factorization (r).

OBJECTIVE

In this work, a new metric for rank selection is proposed based on the elbow method, which was methodically compared against the cophenetic metric.

METHODS

To decide the optimum number rank (r), this study focused on the unit invariant knee (UIK) method of the NMF on gene expression data sets. Since the UIK method requires an extremum distance estimator that is eventually employed for inflection and identification of a knee point, the proposed method finds the first inflection point of the curvature of the residual sum of squares of the proposed algorithms using the UIK method on gene expression data sets as a target matrix.

RESULTS

Computation was conducted for the UIK task using gene expression data of acute lymphoblastic leukemia and acute myeloid leukemia samples. Consequently, the distinct results of NMF were subjected to comparison on different algorithms. The proposed UIK method is easy to perform, fast, free of a priori rank value input, and does not require initial parameters that significantly influence the model's functionality.

CONCLUSIONS

This study demonstrates that the elbow method provides a credible prediction for both gene expression data and for precisely estimating simulated mutational processes data with known dimensions. The proposed UIK method is faster than conventional methods, including metrics utilizing the consensus matrix as a criterion for rank selection, while achieving significantly better computational efficiency without visual inspection on the curvatives. Finally, the suggested rank tuning method based on the elbow method for gene expression data is arguably theoretically superior to the cophenetic measure.

Deep speckle reassignment: towards bootstrapped imaging in complex scattering states with limited speckle grains

Optical imaging through scattering media is a practical challenge with crucial applications in many fields. Many computational imaging methods have been designed for object reconstruction through opaque scattering layers, and remarkable recovery results have been demonstrated in the physical models or learning models. However, most of the imaging approaches are dependent on relatively ideal states with a sufficient number of speckle grains and adequate data volume. Here, the in-depth information with limited speckle grains has been unearthed with speckle reassignment and a bootstrapped imaging method is proposed for reconstruction in complex scattering states. Benefiting from the bootstrap priors-informed data augmentation strategy with a limited training dataset, the validity of the physics-aware learning method has been demonstrated and the high-fidelity reconstruction results through unknown diffusers are obtained. This bootstrapped imaging method with limited speckle grains broadens the way to highly scalable imaging in complex scattering scenes and gives a heuristic reference to practical imaging problems.

Multi-target object scattering imaging with intensity correlation of structured illumination

Imaging through scattering layers based on the optical memory effect (OME) concept has been widely investigated in recent years. Among many scattering scenarios, it is very important to recover hidden targets with proper spatial distribution in the scene where multiple targets out of the OME range exist. In this Letter, we put forward a method for multi-target object scattering imaging. With the help of intensity correlation between the structured illumination patterns and recorded speckle images, the relative position of all hidden targets can be obtained and the movement of the targets within the OME range can be tracked. We experimentally implement scattering imaging with 16 targets and the motion tracking of them. Our results present a significant advance in a large field of view scattering imaging with multiple targets.

Functional imaging through scattering medium via fluorescence speckle demixing and localization

Recently, fluorescence-based optical techniques have emerged as a powerful tool to probe information in the mammalian brain. However, tissue heterogeneities prevent clear imaging of deep neuron bodies due to light scattering. While several up-to-date approaches based on ballistic light allow to retrieve information at shallow depths inside the brain, non-invasive localization and functional imaging at depth still remains a challenge. It was recently shown that functional signals from time-varying fluorescent emitters located behind scattering samples could be retrieved by using a matrix factorization algorithm. Here we show that the seemingly information-less, low-contrast fluorescent speckle patterns recovered by the algorithm can be used to locate each individual emitter, even in the presence of background fluorescence. We test our approach by imaging the temporal activity of large groups of fluorescent sources behind different scattering phantoms mimicking biological tissues, and through a brain slice with a thickness of ~200 micron.

Physics-based neural network for non-invasive control of coherent light in scattering media

Optical imaging through complex media, such as biological tissues or fog, is challenging due to light scattering. In the multiple scattering regime, wavefront shaping provides an effective method to retrieve information; it relies on measuring how the propagation of different optical wavefronts are impacted by scattering. Based on this principle, several wavefront shaping techniques were successfully developed, but most of them are highly invasive and limited to proof-of-principle experiments. Here, we propose to use a neural network approach to non-invasively characterize and control light scattering inside the medium and also to retrieve information of hidden objects buried within it. Unlike most of the recently-proposed approaches, the architecture of our neural network with its layers, connected nodes and activation functions has a true physical meaning as it mimics the propagation of light in our optical system. It is trained with an experimentally-measured input/output dataset built from a series of incident light patterns and corresponding camera snapshots. We apply our physics-based neural network to a fluorescence microscope in epi-configuration and demonstrate its performance through numerical simulations and experiments. This flexible method can include physical priors and we show that it can be applied to other systems as, for example, non-linear or coherent contrast mechanisms.

Large field-of-view non-invasive imaging through scattering layers using fluctuating random illumination

Non-invasive optical imaging techniques are essential diagnostic tools in many fields. Although various recent methods have been proposed to utilize and control light in multiple scattering media, non-invasive optical imaging through and inside scattering layers across a large field of view remains elusive due to the physical limits set by the optical memory effect, especially without wavefront shaping techniques. Here, we demonstrate an approach that enables non-invasive fluorescence imaging behind scattering layers with field-of-views extending well beyond the optical memory effect. The method consists in demixing the speckle patterns emitted by a fluorescent object under variable unknown random illumination, using matrix factorization and a novel fingerprint-based reconstruction. Experimental validation shows the efficiency and robustness of the method with various fluorescent samples, covering a field of view up to three times the optical memory effect range. Our non-invasive imaging technique is simple, neither requires a spatial light modulator nor a guide star, and can be generalized to a wide range of incoherent contrast mechanisms and illumination schemes.