This PDF file contains the front matter associated with SPIE Proceedings Volume 11649, including the Title Page, Copyright information, and Table of Contents.

We present a Computational Miniature Mesoscope that enables 3D fluorescence imaging across an 8-mm field-of-view and 2.5-mm depth-of-field in a single shot, achieving 7-micrometer lateral resolution and better than 200-micrometer axial resolution. The mesoscope has a compact design that integrates a microlens array for imaging and an LED array for excitation on a single platform. Its expanded imaging capability is enabled by computational imaging. We experimentally validate the mesoscopic 3D imaging capability on volumetrically distributed fluorescent beads and fibers. We further quantify the effects of bulk scattering and background fluorescence on phantom experiments.

Fӧrster resonance energy transfer (FRET) based cAMP sensors have become a standard approach to visualize cAMP signals in cells. However, FRET sensors inherently offer low signal-to-noise ratios, making cAMP measurements challenging, especially in 3 spatial dimensions. In previous studies, we used 4-dimensional (x, y, z, and λ) hyperspectral imaging and analysis approaches to measure agonist – induced distribution of cAMP signals in cultured pulmonary microvascular endothelial cells. In many cell types including PMVECs, transient transfection of FRET sensors results in a low transfection efficiency (1-5%), which limits the possibilities of studying spatial distribution of FRET signals in many cells simultaneously. We developed a transgenic rat model that expresses the H187 FRET sensor. In current studies, we utilize hyperspectral imaging and analysis approaches to characterize the distribution of FRET signals in different tissues and vessels harvested from transgenic rats. Hyperspectral z-stacks of tissues and vessels were acquired using a Nikon A1R confocal microscope equipped with 20X multi-immersion objective and a 32 channel PMT detector. Samples were excited using 405 nm and 561 nm lasers and emission was collected from 424 nm -724 nm at 10 nm intervals. Tissues including lungs, heart, mesentery, kidney, expressed the H187 FRET sensor. In conclusion, transgenic rats provide a platform to visualize cAMP signals in-vitro and in-situ.

Multispectral Fluorescence Lifetime Imaging Microscopy (FLIM) is a fundamental tool to study multifold processes in biology and material science. The growing demand for acquisition time reduction requires the parallel acquisition of a multi-dimensional dataset and the exploitation of compressive sensing techniques. In this work we present a multispectral FLIM set-up based on wide-field structured illumination coupled with a spectrometer and a novel time-resolved parallel 18x1 SPAD array detector, working in a single pixel camera scheme. We show the system characterization and its imaging properties varying the compression ratio.

We present a single-shot multiplane widefield imaging strategy using a z-splitter prism, which can be assembled from off-the-self components and only requires a single camera. We further introduce a novel extended-volume 3D deconvolution strategy to suppress far-out-of-focus fluorescence background to significantly improve the contrast of our recorded images, conferring to our system a capacity for quasi optical sectioning. By swapping in different z-splitter configurations, we can prioritize high speed or large 3D field-of-view imaging depending on the application of interest. Moreover, our system can be readily applied to a variety of imaging modalities in addition to fluorescence, such as phase-contrast and darkfield imaging, making it a versatile tool for a wide range of biological or biomedical imaging applications.

Fluorescence Lifetime Imaging Microscopy (FLIM), providing unique quantitative functional information, has gained popularity in various biomedical and molecular biology studies. Here we present an open-source Python package, FlimTK, a toolkit that enables state-of-the-art functions for FLIM image analysis and visualization. It contains comprehensive functionalities for reading FLIM raw files, fluorescence lifetime estimation, heterogeneity analysis, and spatial distribution analysis. FlimTK package is optimized for high performance and ease of use for integration into custom Python-based analysis workflows. FlimTK source code, demo analysis workflows, and tutorial documentation are available for download from GitHub.

Adaptive optics (AO) is used to correct aberrations when focussing deep into specimens and is particularly important in super-resolution nanoscopy methods, which are particularly sensitive. Novel AO methods have been developed to deal with the particular challenges posed by super-resolution microscopes. We show how various three-dimensional nanoscopy applications can be facilitated using AO. Specifically, this includes whole cell and tissue imaging using a 4Pi single molecule localisation microscope, which uses dual opposing objective lenses and two deformable mirrors for improved z resolution; structured illumination microscopy, with adaptive illumination and aberration correction; and STED-fluorescence correlation spectroscopy in living cells.

Fourier ring correlation was recently proposed in fluorescence microscopy as an objective measure of image resolution. In this study we show that FRC has serious limitations. For several common image processing techniques, we show that FRC can be made to be systematically biased to classify objectively lower resolution images as better resolved than objectively higher resolution images. Thus, rather than providing an objective, realistic measure of spatial resolution, care must be taken in interpreting FRC-based results. In revealing some of FRC’s limitations, this study also highlights the need for a more objective, principled approach to estimating resolution.

Breakthroughs in the field of chemistry have enabled surpassing the classical optical diffraction limit by utilizing photo-activated fluorescent molecules. In the single-molecule localization microscopy (SMLM) approach, a sequence of diffraction-limited images, produced by a sparse set of emitting fluorophores with minimally overlapping point- spread functions is acquired, allowing the emitters to be localized with high precision by simple post-processing. However, the low emitter density concept requires lengthy imaging times to achieve full coverage of the imaged specimen on the one hand, and minimal overlap on the other. Thus, this concept in its classical form has low temporal resolution, limiting its application to slow-changing specimens. In recent years, a variety of approaches have been suggested to reduce imaging times by allowing the use of higher emitter densities. One of these methods is the sparsity-based approach for super-resolution microscopy from correlation information of high emitter-density frames, dubbed SPARCOM, which utilizes sparsity in the correlation domain while assuming that the blinking emitters are uncorrelated over time and space, yielding both high temporal and spatial resolution. However, SPARCOM has only been formulated for the two-dimensional setting, where the sample is assumed to be an infinitely thin single-layer, and thus is unsuitable to most biological specimens. In this work, we present an extension of SPARCOM to the more challenging three-dimensional scenario, where we recover a volume from a set of recorded frames, rather than an image.

For the example of digital holographic microscopy (DHM) we explored strategies to discriminate adherent and suspended single cells utilizing biophysical parameters retrieved label-free from DHM quantitative phase images in combination with machine learning (ML). Quantitative DHM phase contrast images of adherent cells were segmented while suspended single cells were analyzed based on a two-dimensional fitting approach. The retrieved parameter clouds were subsequently evaluated with different ML algorithms with the aim of an intuitive and user-friendly data representation. The results of the study demonstrate that our approach is capable for reliable discrimination between different cell types and to distinguish between different phenotypes.

We present a gigapixel-scale multi-aperture microscope capable of measuring a sample’s 3D height profile over multi-centimeter-scale fields of view with a series of single synchronized camera snapshots. Exploiting the overlap redundancy in our multi-aperture camera array microscope, we developed a novel, end-to-end photogrammetric reconstruction algorithm that simultaneously calibrates the cameras’ 3D positions and poses, stitches the acquired images, and generates a coregistered, pixel-wise 3D height map of the sample. Our work opens the door to video-rate 3D monitoring of dynamic scenes at micrometer-scale resolutions and centimeter-scale fields of view.

We formulate a four-dimensional (4-D, space-time dimension) image formation theory of all laser microscopy and optical coherence tomography (OCT) by using 4-D Fourier optics. We define a Fourier transform pair: a 4-D amplitude spread function in space-time domain and a 4-D aperture in spatiotemporal frequency domain. To calculate the 4-D aperture, we also define 4-D pupil functions that include the information on light source spectra in addition to NAs of excitation and detection systems in microscopy. The 4-D aperture is a new concept indicating the sensitivity to object frequencies that contribute to the image formation in microscopy including OCT.

Unwanted background fluorescence in microscopy can occur when light from fluorescent structures scatters against nearby tissues. Bright cell bodies produce haze that mask nearby dim structures, including neuronal fibers. We have developed a method to eliminate this haze by fitting it to a simple function. This method enables clearer post-imaging visualization of axons in our in vivo imaging of neurons. We are investigating ways to increase the computation speed of our background fitting technique to allow for real-time image improvement. The implementation of our method in real-time imaging will broaden the potential applications of this fitting method.

Optical diffraction tomography allows retrieving the 3D refractive index in a non-invasive and label-free manner. A sample is illuminated from various angles and the intensity of the diffracted light is recorded. The light wave can be calculated layer after layer and the inverse problem is usually solved using a gradient descent based algorithm. Here we propose a solution to solve the inverse problem using a neural network where the weights of each layer are the unknown refractive index of the object. Importantly, the matrix product between each layers is replaced by the physics of light propagation.

We report a deep learning-based volumetric imaging framework that uses sparse 2D-scans captured by standard wide-field fluorescence microscopy at arbitrary axial positions within the sample. Through the design of a recurrent neural network, the information from different input planes is blended, and virtually propagated in space to rapidly reconstruct the sample volume over an extended axial range. We validated this deep-learning-based volumetric imaging framework using C. Elegans and nanobead samples to demonstrate a 30-fold reduction in the number of required scans. This versatile and rapid volumetric imaging technique reduces the photon dose on the sample and improves the temporal resolution.

Open-top light-sheet (OTLS) microscopy has recently been developed as a high-throughput, easy-to-use 3D microscopy technique for large specimens [1-2]. The oblique angle of the optical beams relative to the sample plate introduces challenges however, and previous solutions (such as a solid immersion lens) have been limiting [1]. Therefore we have developed a solid immersion meniscus lens (SIMlens), which enables optical beams to transition from air to a higher-index immersion medium without introducing aberrations [3]. A SIMlens is compatible with a turret of air objectives, enabling efficient multi-resolution workflows [4]. We present the first multi-resolution OTLS microscope, enabled by a SIMlens.

We show an integrated three-dimensional (3D) imaging device to acquire volumetric object features through a single exposure on the camera sensor. The system uses a single piece of randomly positioned microlens array as a multi-channel light encoder and a hybrid 3D reconstruction algorithm to achieve compact device footprint and low computational complexity. We validate the system in both simulation models and experiments with resolution targets and fluorescent samples. Our system is capable of volumetric imaging across a volume of 1800 mm3 with a high space-frequency bandwidth. It opens new avenues for miniaturized and implantable imaging device for biomedical applications.

In a tunable 3D structured illumination microscopy (3D-SIM) based on an illumination system comprised by a multi-slit array and a Fresnel biprism, the 3D structured illumination (SI) pattern depends on the design of the slit array. Previous work studied the impact of the illumination pattern design on the achieved extension of the compact support of the tunable 3D-SIM optical transfer function. In this contribution, we use simulations with different illumination designs that utilize a different number of slits, to evaluate system performance by investigating the lateral and axial resolution achieved in the restored images computed with a regularized iterative 3D model-based algorithm, which minimizes the mean square error between the model and data. Three illumination designs are investigated in which the number of slits is N = 3, 9 and 11 while the corresponding distance between the slits is x₀ = 488 μm, 122 μm and 100 μm, respectively. A 3D star-like object, commonly used in resolution analysis of imaging systems, is used in the simulations. Results are quantified using the mean square error and the structured similarity index as well as intensity profiles that show the achieved resolution and robustness to noise in each case. The lateral and axial 3D-SIM theoretical resolution limits are achieved in the presence of 20-dB Poisson noise while small differences are evident in the results from 15-dB simulated data obtained using the 3 investigated designs.

We report a fast and effective lens-free imaging platform for optical diffraction tomography (ODT). Using single wavelength illumination from only 4 angular directions and a lensless inline holographic imaging setup to directly capture the resulting diffraction patterns, our method can reconstruct high quality 3D images of biological samples at micron-scale resolution across a cubic-millimeter-level volume with a compact, scalable and inexpensive system. To achieve this, we developed a compressive tomographic reconstruction algorithm to solve the inverse problem of lens-free ODT by combining Wirtinger derivatives and primal-dual splitting. This fast and inexpensive lens-free tomographic microscopy system provides a promising 3D imaging approach for high throughput biomedical applications.

Light-sheet microscopy has emerged as the preferred means for high-throughput volumetric imaging of cleared tissues. However, there is a need for a user-friendly system that can address diverse imaging applications with varied requirements in terms of resolution (mesoscopic to sub-micron), sample geometry (size, shape, and number), and compatibility with tissue-clearing protocols of different refractive indices. We present a hybrid system that combines a novel non-orthogonal dual-objective and conventional open-top light-sheet architecture for highly versatile multi-scale volumetric imaging.

Multifocal microscopes (MFMs) are becoming increasingly popular in fluorescence microscopy due to their high speed three-dimensional (3D) imaging capabilities. Conventional MFMs use a fixed fabricated grating as the multifocal grating but these are limited to a restricted wavelength range and a fixed object-plane separation. Spatial light modulators (SLMs) represent an alternative to fabricated gratings due to their real-time programmability, providing complete control over emission wavelength range and object plane separations. However, algorithms commonly used to obtain multifocal grating patterns which provide uniform intensity across the subimages are not directly applicable to SLM-based MFMs due to inherent pixel-to-pixel crosstalk effects present in the SLM chip. We recently developed an in-situ iterative algorithm which generates grating patterns that provide near-uniform illumination of the subimages in SLM-based MFMs. This algorithm is universal across wavelengths, object-plane separations, and SLM manufacturers. As part of our efforts to develop an SLM-based MFM that can respond rapidly to changing experimental parameters, we implement a gradient descent-based optimization method. We evaluate its performance in comparison with a grid search based routine. Experimental results obtained on a custom-made SLM-based MFM indicate that the grid-search optimized grating patterns provide superior subimage intensity uniformity versus the gradient-descent method. These experiments also provide an insight into the energy landscape involved in these optimizations. This study increases the utility of SLM-based MFMs in high-speed imaging.

Confocal scanners are used to scan through a large sample depth necessary in both high throughput and high content screening applications. However, this scanning takes time and increases the cost and optical complexity of medical diagnostic devices. Here we present extended depth-of-field engineered point spread function (ePSF) technology, combined with our universal modular subsystem SPINDLE®, to enable precision extended-depth imaging that can be combined with high-content analysis (HCA), without the need for z scanning, and without trading off light or lateral resolution.

Light-sheet microscopy (LSM) is a powerful technique for rapid volumetric imaging of optically cleared specimens. Given the range of possible LSM configurations, designers would benefit from a systematic evaluation of the tradeoffs between different architectures. We present a simulation-based analysis of single- and dual-objective LSM designs for open-top, cleared-tissue imaging. System resolution and contrast are evaluated as functions of the crossing angle between the illumination and collection beams and each beam’s numerical aperture (NA). Our analysis reveals several key tradeoffs to guide designers in addition to potential advantages of a non-orthogonal dual-objective (NODO) architecture for moderate resolution imaging.

Stray light, including scattered and out-of-focus light, can obscure the imaging of dim structures. We added a spatial light modulator at the field stop of our widefield fluorescence microscope to spatially control the illumination and are pursuing several approaches to reduce stray light. We demonstrate confocal capability and a combined >50x signal-to-background ratio improvement by using existing and novel techniques in illumination modulation and image postprocessing. Our approaches offer a simple and low-cost method for adapting existing microscopes to greatly improve the visibility of dim structures that are obscured by bright neighbors.

We present a super-resolving orientation microscopy method for single-particle imaging. In this method, the fluorescence signal from particles is imaged to obtain orientation and localization information simultaneously with higher throughput than photon counting methods. We demonstrate the technique by resolving nanocrystal quantum dot orientations within small clusters. The effects of orientation on coupling efficiency for energy transfer is investigated.

Collagen is one of the most important proteins in mammals, conforming most animal tissues. This work explores how a basic collagen monomer unit is visualized using fluorescence microscopy and how its spatial orientation is determined. Defining the orientation of collagen monomers is not a trivial problem, as the particle has a weak contrast and is relatively small. Possible attach fluorescence tags for contrast, but the size is still a problem for detecting orientation using fluorescence microscopy. This document presents a simulation of the visualization of collagen monomers and two methods for detecting monomer and classifying its orientation. A modify Gabor filter set, and an automatic classifier, trained by convolutional neuronal network (CNN), were used. By evaluating the performance of these two approaches compare to human observation, our results show that it is possible to determine the location and orientation of a single monomer with fluorescence microscopy. These findings can contribute to understanding collagen elements as collagen fibril.

Optical fluorescence microscopy has been widely used to image neurons in the brain. When imaging deep inside of the brain, image signal-to-noise ratio (SNR) severely degrades by scattering, caused by the heterogeneity of different structures. Knowing the refractive index (RI) of the structures can reduce scattering effects and improve image SNR. Currently, most of RI imaging is based on coherent or partial coherent measurements of the transmitted scattered light field, whereas fluorescence microscopy for brain imaging only records incoherent fluorescence signal propagating back to the objective lens in epi-mode. Here, we propose a model to calculate the RI of the 3D scattering tissue only from fluorescence images in reflective geometry. Our method not only successfully reconstructed the RI of a 3D phantom, but also acquired fluorescence images and the unlabeled structures in the same volume simultaneously.

Tomographic inspection of fluorescent labels distributed within a specimen is an important aspect in biology. Light sheet fluorescent microscopy (LSFM) offers a powerful and simple tool to selectively slice the sample and let us directly obtain a tomographic view of the specimen. However, due to non-isotropic resolution of the technique along the axial scanning, one may want to combine different views of the object and add deconvolution to the process in order to achieve higher resolution. Typically, multi-view Bayesian methods based on Richardson-Lucy deconvolution are used for this task once the datasets are exactly registered against each other. In this work, instead, we begin to investigate how to avoid the alignment procedure and use a direct algorithm to form a multi-view tomographic reconstruction. To do this, we developed a new framework based on auto-correlation analysis that let us achieve deconvolved reconstructions starting from blurred auto-correlations. Since the latter are insensitive to shifts, we can combine the auto-correlations coming from multi-view acquisitions without taking care of the registration procedure.

Cerebral microbleeds (CMB) are deposits of blood that accumulate within the brain. An increase in CMBs is associated with an increased risk of cognitive impairment and stroke. The types of vessels associated with CMB formation remains unclear. We recently demonstrated the combined use of exogenous labels, vessel painting, and optical clearing to achieve three-dimensional views of blood vessels and CMBs. Here, we aimed to characterize brain vasculature by quantifying key vasculature-related metrics. An automated algorithm was developed to segment blood vessels within fluorescence images. An open-source neuron tracing software, neuTube, was then used to quantify blood vessel diameters.

Vagus nerve stimulation (VNS) is a method to treat drug-resistant epilepsy and depression, but therapeutic outcomes are often not ideal. Newer electrode designs such as intra-fascicular electrodes offer potential improvements in reducing off-target effects but require a detailed understanding of the fascicular anatomy of the vagus nerve. We have adapted a section-and-image technique, cryo-imaging, with UV excitation to visualize fascicles along the length of the vagus nerve. In addition to offering optical sectioning at the surface via reduced penetration depth, UV illumination also produces sufficient contrast between fascicular structures and connective tissue. Here we demonstrate the utility of this approach in pilot experiments. We imaged fixed, cadaver vagus nerve samples, segmented fascicles, and demonstrated 3D tracking of fascicles. Such data can serve as input for computer models of vagus nerve stimulation.

An efficient and automated image analysis pipeline is essential for extracting quantitative information from multimodal image datasets. In this study, a multimodal optical imaging platform was used to capture CARS, 2PF, and FLIM images from control and drug-treated cells. Images were collected using both fluorescent label-based and label-free approaches. Here we present a single-cell analysis pipeline for the multimodal cellular image analysis. The results demonstrate the capability of our single-cell analysis pipeline for quantitatively measuring the intracellular drug distribution and its longitudinal uptake using a multimodal optical imaging platform, which can provide novel insights into the uptake pathways and target-sites.

We propose a method to construct a three-dimensional volumetric representation of uterine tissue through the full thickness of the uterine wall using Optical Coherence Tomography (OCT). Uterine tissue blocks from human donors were cut into ~4 mm thin slices and imaged on both sides using a commercial OCT system with 2 mm imaging depth. Attenuation compensation, rigid registration, image blending techniques were used to combine the OCT mosaic volumes from each side of the tissue into a single volume. Initial results suggest this method yields a representative 3D view of uterine collagen fiber architecture through the full tissue thickness.

We present Jolab: an open source package for performing full-wave simulation of light propagation in optical systems. Jolab enables a very broad range or researchers, engineers and practitioners to simulate light propagation through complex optical systems. Jolab takes a relatively simple script as its input in which the optical system is defined and light is propagated by each optical components sequentially using built-in functions. Jolab scripts are simple and readable and their structure is designed to mimic the design of optical systems making it easy to learn. We will present a range of examples including time-domain simulations and wavefront-shaping experiments.

In recent years, confocal fluorescence microscopes have become widespread and many Z-stack data have been collected. However, it is the current situation that there is hardly a method for three-dimensional observation of huge amounts of data with sufficient resolution and without loss of information. In light of this situation, we have developed an image processing algorithm, referred to as arbitrary blur algorithm (ABA), which enables post-adjustment of DOF and 3D image observation from a free viewpoint, with high resolution and without loss of information. In this report, we first explain the principle of image processing that enables depth of field (DOF) adjustment and 3D image observation from a free viewpoint using Z-stack data. Next, the results of the simulation are introduced. In the simulation, we show the comparison result of the point image by MIP (maximum intensity projection) method and the point image of the proposed method and clarify the superiority of the proposed method. Next, we introduce the results of the experimental verification. For experimental verification, we show the results using wide field fluorescence microscope Z-stack data and confocal fluorescence microscope Z-stack data. As confocal data, observation examples of mouse hippocampal neuron synapses are shown in comparison with the results displayed by existing software.

Scanning microscopy’s resolution of highly localized dynamics in awake animals is limited by bulk motion. We present a modular optical flow sensor based on spectral domain OCT which measures the specimen’s 3D displacement in real-time. A rose-function scan pattern rapidly acquires B-lines used to estimate displacements via an adaptive cross-correlation approach. Axial displacements are estimated from phase fluctuations apparent in the cross-correlation of the complex SD-OCT signal. The technique’s 3-dimensional readout rate and sensitivity to physiologically-relevant motion frequencies and magnitudes is evaluated.

Light-sheet microscopy (LSM) has emerged as the technique of choice for many biologists imaging large cleared tissues due to its speed and optical efficiency, which make it possible to generate massive datasets of large specimens at high resolution. Here, we build on several recent innovations in LSM to present a non-orthogonal dual-objective (NODO) LSM system with axial sweeping in an open-top configuration. This system is specifically designed to image large cleared brain tissues, such as for axonal connectomics, and provides subcellular resolution (0.3 µm lateral, 2 µm axial) of large cleared samples up to 8 mm thick.

F¨orster resonance energy transfer (FRET) is a tool used for studying various biological process as well as for measuring molecular distances. This process can occur when the emission spectrum of the donor fluorophore overlaps with the excitation spectrum of the acceptor, and the fluorophores are in close enough distance for the energy to pass from donor to acceptor non-radiatively. The efficiency of this energy transfer is dependent on the distance and orientation of the fluorophores, in addition to their overlapping spectra. Here we present a study to assess the impact of tissue autofluorescence on estimates of FRET efficiency and fluorophore abundance within experimental cellular images in tissue. To accomplish this we performed a theoretical sensitivity analysis on FRET rat kidney control images with varying ranges of donor and acceptor fluorophores to observe their pixel by pixel responses. In the experimental data, the donor was the Turquoise fluorescent protein and the acceptor fluorophore was the Venus fluorescent protein. Detection of the acceptor was more difficult due to its excitation spectrum closely resembling the autofluorescence spectrum from the base image while the emission spectrum of the Turquoise donor was more unique and easier to detect. In addition, variable FRET efficiencies were added to data at different fluorophore levels to compare the visible abundance of FRET to the autofluorescence in the resultant images. This analysis can benefit future work by furthering the understanding of the donor and acceptor concentrations needed for strong FRET measurements.

Efficient accurate Gaussian fitting is an important topic in many applications, take localization based super-resolution microscopy and image scanning microscopy for example, which requires localizing plenty of Gaussian patterns for accurately reconstructing a super-resolution image. Existing Gaussian fitting methods usually require inputting a good initial value for all parameters for efficient and robust fitting, which apparently is not suitable for the task of large scale Gaussian fitting with a computer. It would be even more challenge to estimate an appropriate initial value for all parameters and guarantee the robustness of the fitting algorithm for low signal-to-noise ratio measured data with strong background. In this paper, we propose a two-step fitting algorithm for robust and accurate 2D Gaussian fitting without inputting any initial parameter value. Our simulation shows that the performance of the fitting algorithm can be improved significantly with the proposed parameter initial value estimation algorithm.

Photoacoustic imaging is becoming a very promising tool for the research of living organisms. It combines the high contrast of optical imaging and the high resolution of acoustic imaging to realize the imaging of absorption clusters in biological tissues. Since the scattering of ultrasound signals in biological tissues is 2-3 orders of magnitude weaker than the scattering of light in biological tissues, the endogenous absorption difference of tissues is directly used in the imaging process, so photoacoustic imaging has the advantages of deep imaging depth and non-destructive. As an important branch of photoacoustic imaging, photoacoustic microscopy can provide micron-level or even sub-micron-level imaging resolution, which is of great significance for biological research such as blood vessel detection. Since the lateral resolution of the photoacoustic microscopy imaging system depends on the focus of the laser, a higher resolution can be obtained by increasing the numerical aperture of the condenser objective. However, a large numerical aperture usually means a shorter working distance and makes the entire imaging system very sensitive to small optical defects. Therefore, the improvement of resolution through this method will be limited in practical applications. This paper implements a method of using iterative deconvolution to obtain a high-resolution photoacoustic image of the brain. The focal spot of the photoacoustic microscopy is measured to obtain the lateral PSF (point spread function) of the system. Making the measured PSF as the initial system PSF to perform Lucy- Richardson (LR) deconvolution. The image resolution of cerebral vasculature obtained by this method is higher. The full width at half maximum (FWHM) of width of the same cerebral capillaries before and after deconvolution are 7 μm and 3.6 μm, respectively, and the image definition is increased by about 1.9 times. Experiments show that this method can further improve the clarity of photoacoustic images of cerebral capillaries, which lays the foundation for further research on brain imaging.