This PDF file contains the front matter associated with SPIE Proceedings Volume 10871 including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

Spectral domain optical coherence tomography (SD-OCT) and photoacoustic microscopy (PAM) were integrated to improve the visualization of retinal vein occlusion (RVO) and retinal neovascularization (RNV). RVO and RNV were observed in living, New Zealand rabbits by Rose Bengal laser-induced retinal vein thrombosis. Multimodal imaging techniques including PAM, OCT, fluorescein angiography (FA), and color fundus photographs were utilized to observe and analyze changes in the retinal vasculature. Spectral domain OCT identified RNV cross-sectional structures, and progressive changes in retinal anatomy due to angiogenesis from the photothermal treatment were monitored at 4, 28, 35, 49, and 90 days post-laser. Compared to alternative methods, PAM in vivo high-contrast imaging possesses the capabilities to accurately visualize the treatment margins of occluded vasculature and areas of RNV. Results were obtained using a laser energy of 80 nJ, which is half a dose below the American National Standards Institute safety limit. The PAM system also demonstrated an increased depth of penetration, which provided high resolution images of the choroid and retinal vasculature when the optical absorption of hemoglobin was used to improve the visualize blood flow. Current modalities of imaging possess the ability to visualize RNV and RVO in two-dimensional and three-dimensional angiography. However, the system integrating both PAM and OCT can better visualize the depth and position of microvessels as well as surrounding structures. This specific multimodal ocular imaging technique may therefore be an improved technology to document minute changes in rabbit eye retinal vasculature while proving to be both safe and efficient.

Laser therapy has been used to perform both ablation and coagulation of diseased tissue. To avoid over or under exposure, monitoring such therapies with a cost-effective method remains an issue however. We present an integrated solution based on an optical coherence tomography (OCT) system allowing simultaneous imaging, quantitative monitoring and therapy delivery in real-time. The system exploits a double-clad fiber coupler (DCFC) to inject the OCT signal into the double-clad fiber (DCF) core and the therapy laser into the inner cladding making them co-localized. The single fiber solution permits both imaging and therapy at the same time. Furthermore, the DCFC allows the implementation of our technique in any OCT system sharing the same wavelength bandwidth. Therapy monitoring is achieved by measuring the speckle intensity decorrelation. During coagulation, the optical properties of the tissue start to vary, thereby changing the speckle intensity pattern seen in the OCT tomograms. The proposed algorithm includes both novel motion and noise corrections, extending the usable monitoring depth. Furthermore, the code has been optimized to run during therapy providing real-time monitoring. In a proof of concept experiment, a system was built with a 532 nm CW laser for therapy and a 1310 nm swept-source laser for OCT imaging. We present ex-vivo cross-sectional imaging and monitoring during therapy. Experimental results were validated against Monte-Carlo simulations and visual inspection.

Hyperspectral imaging (HSI) is a spectroscopic technique which captures images at a high contrast over a wide range of wavelengths to show pixel specific composition. Traditional uses of HSI include: satellite imagery, food distribution quality control and digital archaeological reconstruction. Our lab has focused on developing applications of HSI fluorescence imaging systems to study molecule-specific detection for rapid cell signaling events or real-time endoscopic screening. Previously, we have developed a prototype spectral light source, using our modified imaging technique, excitationscanning hyperspectral imaging (HIFEX), coupled to a commercial colonoscope for feasibility testing. The 16 wavelength LED array was combined, using a multi-branched solid light guide, to couple to the scope’s optical input. The prototype acquired a spectral scan at near video-rate speeds (~8 fps). The prototype could operate at very rapid wavelength switch speeds, limited to the on/off rates of the LEDs (~10 μs), but imaging speed was limited due to optical transmission losses (~98%) through the solid light guide. Here we present a continuation of our previous work in performing an in-depth analysis of the solid light guide to optimize the optical intensity throughput. The parameters evaluated include: LED intensity input, geometry (branch curvature and combination) and light propagation using outer claddings. Simulations were conducted using a Monte Carlo ray tracing software (TracePro). Results show that transmission within the branched light guide may be optimized through LED focusing lenses, bend radii and smooth tangential branch merges. Future work will test a new fabricated light guide from the optimized model framework.

Detection of lipid core plaques during coronary catheterization procedures may lead to secondary prevention, reduction in procedural complications, and eventually the development of reliable non-invasive methods for primary and secondary prevention. Prospective studies are in progress to test the hypothesis that combined NIRS-IVUS imaging can detect vulnerable coronary plaques and guide preventive therapy.

Among Central Nervous System (CNS) tumors, diffuse glioma are the most infiltrating and malignant tumors. According to the World Health Organization (WHO), they are classified into different grades, referring to their pathological class and histological properties. To treat glioma tumors, many methods have been proposed, still the standard one remains the maximal safe total resection. During the operation, many difficulties obstruct the surgeon to identify the infiltrated areas, which contains diffuse tumors cells, around the solid area of the tumor, overlapping the healthy areas, and presenting the same visual appearances. If not totally removed, these infiltrating zones can increase the risk of recurrence and affects the survival rate of patient. To overcome this problem, we develop a multimodal two-photon endomicroscope, based on the endogenous fluorescence of brain tissues, to assist the surgeon during the surgery. The tool will provide him information on the infiltrated areas and their histological nature. In this paper, we tried to discriminate between metastasis, low grade and high grade glioma from healthy fresh tissues, presenting a multimodal study using deep ultraviolet, visible and near infrared excitation to acquire spectral measurements, Fluorescence Lifetime Imaging (FLIM) and Two-Photon Emission Fluorescence (TPEF) imaging. We compared also our TPEF and FLIM images to the histological images.

Prostate cancer is the most diagnosed form of cancer among American men and, in vast proportion, the standard of care treatment includes radical prostatectomy. Important risk factors associated with prostatectomies are the presence of post-surgery residual prostate tissue and positive cancer margins, potentially leading to recurrences. Prostate histopathology analysis following the procedure is used to determine follow-up treatment. However, only a limited fraction of the prostate margins can be sampled, which can lead to suboptimal evaluation and treatment. Here we present the development of a wide-field multimodal imaging system designed to quantify intrinsic tissue fluorescence and map scattering and absorption coefficients using spatial frequency domain imaging (SFDI). The system allows targeting of suspicious prostate regions to guide histopathology analysis, aiming to improve diagnostic accuracy and treatment planning. Tissue excitation for endogenous fluorescence is achieved with a 405 nm laser diode and, for SFDI, a digital light projector transmits structured white light used to reconstruct tissue optical properties (absorption, scattering) between 420 and 720 nm. A light transport model-based quantification algorithm then corrects the fluorescence spectra for tissue attenuation, lending a biomarker that correlates with local fluorophore concentrations. Spectral and spatial calibration of both modalities was done on optical phantoms and validation of the fluorescence quantification on biological tissue. Finally, imaging results are presented for 5 human prostates interrogated with the system, along with spatially-registered histopathology analyses. Future work involves massive data acquisition and development of artificial intelligence models for tissue classification (prostate, non-prostate; healthy, cancerous) and adaptation for intraoperative use.

Over the past three decades, non-invasive molecular imaging via optical tomography has garnered attention in the field of preclinical imaging thanks to its high sensitivity and ability to image multiple biomarkers simultaneously. However, it is still very challenging to image intact tissues with high resolution while retaining the two aforementioned characteristics. Over the last few years, our group has pioneered Mesoscopic Fluorescence Molecular Tomography (MFMT), a novel imaging modality that recapitulates the 3D distribution of fluorescent markers within thick and diffuse samples (< 3 mm) with spatial resolution ~100 µm. Still, as a diffuse optical inverse problem, the image formation can be challenging due to its ill-conditioned nature. Herein, we report on the fusion of MFMT with Optical Coherence Tomography (OCT) to provide both structural and molecular imaging capabilities. Moreover, we leverage the OCT information to impart structural priors that facilitate the optical inverse problem in MFMT. We demonstrate the capability and utility of this novel platform on bioprinted tissues, fluorescent polymer letters in agar phantoms, and on microfabricated beads at different imaging depths.

Alterations in tumor microvascular architecture are associated with resistance to several breast cancer therapies, and may be important markers for in vivo detection of resistance. Our goal is to map how micro-scale alterations in tumor vasculature manifest at the tissue level. To this end, we developed a multiscale preclinical imaging technique called Diffuse and Nonlinear Imaging (DNI) that integrates Spatial Frequency Domain Imaging (SFDI) for tissue-level mapping of tumor optical properties and hemodynamics, with Multiphoton Microscopy (MPM) to image tumor microvascular architecture with cellular resolution. Importantly, SFDI measures the same parameters as clinical Diffuse Optical Spectroscopy, providing a pathway to the clinic for microvascular imaging biomarkers. We demonstrated that the dual modality system can be spatially co-registered with high accuracy and precision (≤ 50 µm), and can be matched in optical sampling depth based on wavelength and spatial frequency selection. We also conducted an in vivo DNI study of untreated murine mammary tumors (Py230) in female C57BL/6 mice, and found strong multiscale relationships between tumor oxygen saturation and micro-vessel diameter, as well as deoxyhemoglobin concentrations and micro-vessel length (|Pearson’s ρ| > 0.5, p < 0.05). We carried out in vivo DNI monitoring in two mammary tumor xenograft models grown in BALB/c athymic nude female mice; one model was responsive to Trastuzumab (Herceptin®) (BT474) and the other was resistant (HR6). This presentation will report on characterizing the vascular structure-function relationships with DNI across length scales within each model, and differences in the multiscale vascular relationships between the models.

Although diffuse optical tomography (DOT) is able to obtain valuable functional information, its routine use in clinic is hampered by its poor spatial resolution and quantitative accuracy. Previously, our team introduced Photo-Magnetic Imaging (PMI) to overcome the limitation of DOT. PMI is a hybrid modality that synergistically utilizes optics and Magnetic Resonance Imaging (MRI). While illuminating the imaged medium by near-infrared laser, the induced internal temperature increase is measured using Magnetic Resonance Thermometry (MRT). Using these MRT maps, optical absorption maps at the laser's wavelength can be recovered using the dedicated PMI image reconstruction algorithm. In this paper, we present the result of the first validation simulation study of multi-wavelength PMI that utilizes five different laser wavelengths ranging between 760 and 980 nm. Using the high resolution wavelength specific absorption maps, PMI successfully recovered the concentration of three dyes, used as chromophore in the composition of our phantom, with high spatial resolution and quantitative accuracy. By providing functional information at high resolution, multi-wavelength PMI will be a valuable tool for monitoring tissue physiology, cancer detection and monitoring.

X-ray luminescence computed tomography (XLCT) is an emerging hybrid molecular imaging modality with great promises in overcoming the strong optical scattering in deep tissues for good spatial resolution. Though the narrow x-ray beam XLCT imaging has been demonstrated to obtain high spatial resolution at depth, it suffers from a relatively long measurement time, hindering its practical applications. Recently, we have designed a focused x-ray beam based XLCT imaging system and have successfully performed imaging in about 12.5 minutes per section imaging for a mouse sized object. Following this previous work, in this current study, we have performed XLCT imaging using our focused x-ray beam for both a tissue-mimicking phantom and for the first time, with a euthanized mouse embedded with a capillary tube target filled with 10.0 mg/mL of GOS:Eu³⁺ microphosphors and have shown that the data acquisition time could be reduced substantially to less than 10 milliseconds per linear scan step compared to the previous study which used 1 second per linear step. In addition, the targets were reconstructed with a high location accuracy and good shape. In the current setup, the total measurement time for a mouse sized object could be reduced to about 7.5 seconds per section imaging, a major improvement from previous studies.

Fluorescence lifetime imaging (FLI) has become an invaluable tool in the biomedical field by providing unique, quantitative information about biochemical events and interactions taking place within specimens of interest. Applications of FLI range from superresolution microscopy to whole body imaging using visible and near-infrared fluorophores. However, quantifying lifetime can still be a challenging task especially in the case of bi-exponential applications. In such cases, model based iterative fitting is typically employed but necessitate setting up multiple parameters ad hoc and can be computationally expensive. These facts have limited the universal appeal of the technique and methodologies can be specific to certain applications/technology or laboratory bound. Herein, we propose a novel approach based on Deep Learning (DL) to quantify bi-lifetime Forster Resonance Energy Transfer (FRET). Our deep neural network outputs three images consisting of both lifetimes and fractional amplitude. The network is trained using synthetic data and then validated on experimental FLI microscopic (FLIM) and macroscopic data sets (MFLI). Our results demonstrate that DL is well suited to quantify wide-field bi-exponential fluorescence lifetime accurately and in real time, even when it is difficult to obtain large scale experimental training data.

The limited sampling of biopsy and histopathology can lead to incomplete and/or inaccurate assessment of basal cell carcinomas (BCCs), subtypes and depth, which can affect diagnosis and treatment outcome. Reflectance confocal microscopy (RCM) combined with optical coherence tomography (OCT) can help achieve comprehensive 3-dimensional sampling in vivo, which may improve the diagnostic accuracy and margin assessment of BCCs. In a clinical study, we tested a combined RCM-OCT probe on 85 patients, with either clinically-suspicious (n=60, in intact skin) or biopsy-proven BCCs (n=25, in scarred skin). We correlated BCC features in RCM and OCT images with histopathology, calculated diagnostic accuracy and correlated depth predicted by OCT with histopathologically measured depth. The main features were small tumors extending from the basal cell layer at the dermal-epidermal junction; small and large tumor nests; in dermis; dark silhouettes; dilated blood vessels; horn cyst and bright peritumoral stroma. Deeper features such as necrosis and intratumoral mucin pools were correlated on OCT and histology. Higher sensitivity and negative predictive value (100%) and comparable specificity (48% vs 56% on RCM) and positive predictive value (82.19 vs 84.59 % on RCM) were observed for the combined RCM-OCT device for diagnosis of all lesions (n=85). Relatively higher specificity (94.1%) and positive predictive value (75%) were observed in the clinically suspicious lesions (n=60, in intact skin). High correlation was observed (R=0.86) between the OCT predicted depth and histopathologically measured depth. Therefore, RCM-OCT imaging may be prospectively used to comprehensively diagnose suspicious BCC lesions, determine subtype and triage for treatment.

Colorectal cancer is the second leading cause of cancer deaths in the United States and causes over 50,000 deaths annually. The standard of care for colorectal cancer detection and prevention is an optical colonoscopy and polypectomy. However, over 20% of the polyps are typically missed during a standard colonoscopy procedure and 60% of colorectal cancer cases are attributed to these missed polyps. Surface topography plays a vital role in identification and characterization of lesions, but topographic features often appear subtle to a conventional endoscope. Chromoendoscopy can highlight topographic features of the mucosa and has shown to improve lesion detection rate, but requires dedicated training and increases procedure time. Photometric stereo endoscopy captures this topography but is qualitative due to unknown working distances from each point of mucosa to the endoscope. In this work, we use deep learning to estimate a depth map from an endoscope camera with four alternating light sources. Since endoscopy videos with ground truth depth maps are challenging to attain, we generated synthetic data using graphical rendering from an anatomically realistic 3D colon model and a forward model of a virtual endoscope with alternating light sources. We propose an encoder-decoder style deep network, where the encoder is split into four branches of sub-encoder networks that simultaneously extract features from each of the four sources and fuse these feature maps as the network goes deeper. This is complemented by skip connections, which maintain spatial consistency when the features are decoded. We demonstrate that, when compared to monocular depth estimation, this setup can reduce the average NRMS error for depth estimation in a silicone colon phantom by 38% and in a pig colon by 31%.

In this work we explore the problem of multiclass classification where the classifier may abstain from classifying on some observation. We derivate a new surrogate loss function and a multiclass decision rule by using a reject threshold on posterior probabilities in the Bayes decision rule, known as Chow's rule. The goal of the decision rule is to minimize the value of given misprediction and rejection cost functions specified by the user. We suggest a general training algorithm by plug-in the surrogate loss in to Support Vector Machine (SVM) structure. We then test the algorithm on various real -life problem in the photonic medical sensing field where accuracy is critical. We present an example of a non-invasive way of detecting glucose level in blood to help patients with Diabetes mellitus diseases while the sensing is performed with speckle-based approach to analyze remote sensing of biomedical parameters. The results will show that the value of the reject threshold has importance in determining how many samples to reject and in the overall accuracy of prediction. As the threshold grow so does the number of samples rejected and overall accuracy, meaning that only samples with strong confidence are outputted in the classification process. A very important point in working with reject option is that there is a tradeoff between the number of samples being rejected and the accuracy of the labeled samples. High precision comes with high rejection rate, while low rate of rejection derogates from the general correctness of the output.

Time-domain diffuse optical tomography (TD-DOT) is interested to many scholars due to its unique features such as rich information in each measured pulse, elimination of cross-talk in the reconstruction, and no need for the measurement calibration, although TD-DOT imaging system requires very expensive equipment. Usually, the generalized pulse spectrum technique (GPST) is applied to transform the real domain optical diffusion forward model into its Laplacian domain. However, the application of TD-DOT has been limited by the low spatial resolution due to strong optical scattering in tissues. In this study, we introduced the anatomical guidance from computed tomography (CT) and positron emission tomography (PET) into TD-DOT to improve the quality of the reconstructed TD-DOT images with the soft prior method. Furthermore, we have investigated to maximize the utilization of the rich information contained in the measured pulses by increasing the frequency number in the Laplacian transformation. To validate the proposed method, we have performed numerical simulation studies by using the CT/PET breast images of a breast cancer patient, in which we generated the breast mesh from CT images and obtained the breast tumor position and size from the PET images. The soft prior guidance based TD-DOT reconstruction was performed in the Laplacian domain. Our simulation results indicate that the spatial resolution and the accuracy of the reconstructed TD-DOT images have been improved substantially after applying the anatomical guidance.

Imaging studies are one of the leading drivers of modern medical decision making, and thus, their accessibility to healthcare providers and patients is of critical importance. However, current techniques for storage and transferring medical imaging data are inconvenient and sometimes wholly inadequate. In this paper, we propose a decentralized autonomous medical image processing approach using blockchain technology. Blockchain will enable the sharing of key relevant data using a distributed, decentralized, shared ledger that is available to participants. We outline a framework that utilizes blockchain to enable users to access imaging data in a secure and autonomous manner. A user case is experimentally investigated to validate our proposed approach.

More than 54 million Americans have or are at high risk of developing a metabolic bone disease; disorders of bone strength that leave individuals with fragile bones and disabilities. The gold standard to evaluate these diseases is dual energy x-ray absorptiometry, but this only measures mineral content. These diseases, however, impact collagen and mineral integrity which impede the bone’s ability to store hormones, proteoglycans, and glycoproteins imperative to homeostasis. We have established a second harmonic generation (SHG) polarimetric assay that describes bone collagen organization. To further our analysis, we propose multimodal optical evaluation of bone quality with third harmonic generation (THG) to measure osteocyte dendritic processes. This method of analysis could be used to evaluate the disease state of bone and response to therapy.

Colorectal cancer accounts for an estimated 8% of cancer deaths in the United States with a five-year survival rate of 55-75%. The early detection and removal of precancerous lesions is critical for reducing mortality, but subtle neoplastic growths, such as non-polypoid lesions, often go undetected during routine colonoscopy. Current approaches to flat or depressed lesion detection are ineffective due to the poor contrast of subtle features in white light endoscopy. Towards improving colorectal lesion contrast, we present an endoscopic light source with custom laser channels for multimodal color, topographic, and speckle contrast flow imaging. Three red-green-blue laser units, paired with laser speckle reducers, are coupled into endoscopic fiber optic light guides in a benchtop endoscope. Tissue phantom topography is reconstructed using alternating illumination of the laser units and a photometric stereo endoscopy algorithm. The contrast of flow regions is enhanced in an optical flow phantom using laser speckle contrast imaging. Further, the system retains the ability to offer white light and narrow band illumination modes with improved power efficiency, a reduced size, and longer lifetimes compared to conventional endoscopic arc lamp sources. This novel endoscopic light source design shows promise for increasing the detection of subtle lesions in routine colonoscopy screening.

Combining optical coherence tomography (OCT) and multiphoton microscopy (MPM) can provide multimodal imaging of the microstructure of biological tissues. As a functional extension of conventional OCT, speckle variance OCT (SVOCT) can be applied to image microvasculature to improve the blood vessel visualization. In this paper, a combined SVOCT and MPM system is developed to visualize the chorioallantoic membrane (CAM) of a chick embryo, which contains extensive blood vessel network. Based on the different temporal decorrelation characteristics of the fluid flow and the surrounding stationary structure, SV-OCT enables enhanced contrast of fluid flow from the surrounding structure. As a result, SV-OCT can achieve detailed mapping of the CAM microvasculature at the tissue level. Meanwhile, MPM enables vascular imaging at the cellular level, where two-photon excitation fluorescence (TPEF) images fluorescein dye injected into the blood stream, and second harmonic generation (SHG) visualizes the collagen fiber structures in the vessel wall and the surrounding tissues. Therefore, the combined SV-OCT and MPM system provides complementary information about the microvasculature structures in the chick CAM. The combined system is shown to be a powerful tool for interpreting the microvasculature, by allowing the visualization of the blood vessel network in a relatively large field of view at the tissue level with SV-OCT, and by providing cellular-level information in local regions of interest with MPM.

Cherenkov luminescence tomography (CLT) has become a novel three-dimensional (3D) non-invasive technology for biomedical applications such as tumor detection, pharmacodynamics evaluation, etc. However, the reconstruction of CLT still remains a challenging task because of the strong absorbing effect and scattering effect of Cherenkov photon transport process. In this study, we proposed a novel robust sparse reconstruction method named look ahead orthogonal matching pursuit (LAOMP) algorithm to improve the robustness and accuracy of reconstruction for CLT instead of traditional OMP algorithm based on a look ahead strategy. To validate the reconstruction performance of LAOMP method, a series of numerical simulations were conducted. The results showed that LAOMP method obtained the higher robustness and accuracy in locating the optical sources compared with the OMP and StOMP algorithms.

Diffuse optical tomography in a transrectal configuration (TR-DOT) has been developed to monitor progression of the photocoagulation front during interstitial photothermal therapy (PTT) of focal prostate cancer. Building on simulations and feasibility studies in coagulating phantoms, ex vivo porcine muscle and ex vivo canine prostate, the technique was tested here in preclinical canine prostate models in vivo to assess the signal stability and device sensitivity for clinical translation. Co-registration of DOT measurements with magnetic resonance imaging (MRI)-based thermometry in photocoagulating phantoms provide temperature profiles for correlation with DOT signals changes. DOT measurements were performed near the treatment fiber tip during PTT with near-infrared light delivery in healthy canine prostate and kidney in vivo. The results were compared with numerical simulations, given the tissue optical absorption and scattering properties. In parallel studies, TR-DOT is being investigated to localize prostate tumor and to enhance PTT using porphyrin-lipid nanoparticles (porphysomes, pPS). The DOT system is being further optimized for real-time monitoring of focal prostate PTT.

Cerenkov fluorescence imaging (CLI) has set a bridge between optical and nuclear imaging technologies by using an optical method to detect the distribution of radiotracers. Combining the emerged CLI technique with a clinical endoscope, the Cerenkov luminescence endoscope (CLE) was developed to avoid the problem of the poor penetration depth of the Cerenkov light. However, due to low energy of the Cerenkov light and the transportation loss during endoscopic imaging, the acquisition time of CLE signal is long and the imaging results are poor, which has limited the clinical applications of CLE. There are two ways to improve the availability of the current CLE system. First is to enhance the emitted signals of the Cerenkov light at the source end by developing new kinds of imaging probes or selecting high yield radionuclides. However, this will introduce the in vivo unfriendly problem in clinical translations. The second method is to improve the detection sensitivity of CLE system by optimizing the structure of the system. Here, we customized four endoscopes with different field of view (FOV) angles of endoscope probe and different monofilament diameters of imaging fiber bundles. By comparing the results obtained by different CLE systems, we optimized the parameters of system. The CLE imaging of 18F-FDG showed that when the distance between the probe and radionuclide source was fixed, smaller angle of FOV and lager monofilament diameter will provide higher collection efficiency.

The heart assist pump of a pneumatic type has a pneumatic chamber and a blood chamber separated by a flaccid membrane. During operation, the membrane changes its shape. Its reconstruction is possible using a camera with a wide-angle lens and the DFD method for visual distance measurement. The measurement is carried out simultaneously at characteristic points indicated by passive markers placed on a membrane surface. Due to their limited number, to obtain a proper numerical description of a membrane shape, spatial interpolation in the actual dimensions is necessary. In the paper, the method of interpolation and the results of reconstruction tests based on 3D printed models were presented.

As a promising tomographic method in preclinical research, fluorescence molecular tomography (FMT) can obtain real-time three-dimensional (3D) visualization for in vivo studies. However, because of the ill-posed nature and sensitivity to noise of the inverse problem, it remains challenging for effective and robust reconstruction of fluorescent probe distribution in animals. In this study, we present a two-stage matching pursuit (TSMP) method. The iterative process is divided into two processes: In the first stage, we iterate several times using the OMP algorithm to improve the accuracy of the support set, which is because most of the atoms selected by the OMP algorithm are accurate. In the second stage, we use CoSaMP algorithm to iterative. The initial input of the second stage is the residual and atom obtained by the first stage OMP algorithm, which can change the dependence of CoSaMP to sparsity. Meanwhile, considering the time of reconstruction, we set the iterative times of the first stage to K/2 (K is the sparisty). Because of the accuracy of the initial output and the choice of atomic criteria, the proposed algorithm has better performance than OMP and CoSaMP algorithm. The result of numerical simulation show that TSMP method can not only achieves accurate and desirable fluorescent source reconstruction, but also demonstrates enhanced robustness to noise.

Computed tomography (CT) imaging acquires patient images using radiation. However, scanning with high doses of radiation can pose a risk to health, because of radiation hazards. Although the risk to the human body during a CT scan can be reduced by reducing the amount of radiation, the quality of the acquired images may deteriorate. Recently, denoising methods using nonlocal means or block matching and 3D filtering were demonstrated to be effective for denoising CT images. These methods performed denoising by adapting to the noise level, according to the position of the image. However, CT images exhibit different magnitudes of noise at different spatial frequencies, as can be observed in their noise power spectrum. Therefore, a method operating in the frequency space, which can accurately model the CT noise and reduce it effectively, is necessary. In this paper, we present a CT denoising method based on edge-preservation segmentation and denoising using mask nonharmonic analysis (mask NHA). Mask NHA can accurately analyze frequencies with high resolutions when applied to the edge preservation area. By using a whitening filter, we provide noise reduction for specific CT noises and improve the image quality of low-dose CT images. A denoising simulation was performed on a standard-dose CT image to which CT noise was added and the performance of the proposed method was compared to that of conventional methods. The proposed method was found to improve the peak signal-to-noise ratio by 3 to 5 dB, compared to the conventional mask NHA.