KEYWORDS: Digital holography, 3D image reconstruction, Reverse modeling, Holograms, Deep learning, Mathematical optimization, Image restoration, Education and training
Untrained Physics-based Deep Learning (DL) methods for digital holography have gained significant attention due to their benefits, such as not requiring an annotated training dataset, and providing interpretability since utilizing the governing laws of hologram formation. However, they are sensitive to the hard-to-obtain precise object distance from the imaging plane, posing the Autofocusing challenge. Conventional solutions involve reconstructing image stacks for different potential distances and applying focus metrics to select the best results, which apparently is computationally inefficient. In contrast, recently developed DL-based methods treat it as a supervised task, which again needs annotated data and lacks generalizability. To address this issue, we propose reverse-attention loss, a weighted sum of losses for all possible candidates with learnable weights. This is a pioneering approach to addressing the Autofocusing challenge in untrained deep-learning methods. Both theoretical analysis and experiments demonstrate its superiority in efficiency and accuracy. Interestingly, our method presents a significant reconstruction performance over rival methods (i.e. alternating descent-like optimization, non-weighted loss integration, and random distance assignment) and even is almost equal to that achieved with a precisely known object distance. For example, the difference is less than 1dB in PSNR and 0.002 in SSIM for the target sample in our experiment.
Histologic examination of skin biopsies is currently the gold standard to definitively diagnose malignant skin lesions; however, biopsies are minor, invasive procedures with potential risks. With the advancement of imaging techniques such as laser speckle contrast imaging (LSCI), it is now possible to evaluate neoplastic skin lesions in real-time and noninvasively. LSCI has been widely used to image surface blood flow in tissues, such as skin, retina, and brain. In this preliminary study, we hypothesized that blood flow within microvessels differs between neoplastic and non-neoplastic skin. This study presents a descriptive demonstration of LSCI application in dermatology. LSCI was utilized to assess surface blood flow in potentially neoplastic skin lesions at our institution’s dermatology clinics. Preliminary data demonstrated decreased contrast within speckle contrast images of malignant and premalignant skin lesions, suggesting increased blood flow to these areas of interest. LSCI may show utility as a noninvasive technique to evaluate neoplastic skin lesions prior to biopsy; however, further systematic optimization is required.
Articular cartilage in the tibiofemoral joint contains unique tissue microstructures that serve specific functions, including reduction of friction and distributing the dynamic and static cyclic loading at the ends of diarthrodial joints. A proficient understanding of these microstructures can lead to significant clinical advances in diagnosing orthopedic diseases such as osteoarthritis and improving cartilage repairs. The surface of tibiofemoral condyles can be roughly separated into loadbearing and meniscus-covered areas. Due to the difference in mechanical loading between the two regions, we hypothesize that their microstructures differ. To test this hypothesis, we used cartilage punches harvested from the tibial condyle of porcine knee joints as an example tissue and a custom nonlinear optical microscope for performing a dye-free imaging study. The custom nonlinear optical microscope could simultaneously acquire Two-Photon excitation Auto-Fluorescence (TPAF) and Second Harmonic Generation (SHG) images. Through the TPAF channel, elastin fibers are visible along with chondrocytes. The SHG channel was utilized for observing the vast collagen network and its evident orientation throughout the tibial condyle. Images were analyzed by ImageJ to reveal alignment angles of the collagen network and elastin fibers. The load-bearing region exhibits a denser uniform collagen network with minimum elastin fibers. In contrast, the meniscuscovered areas have a distinctive collagen orientation with a greater magnitude of co-localized elastin fibers. The biological differences are likely derived from their different biomechanical environments in the tibiofemoral joint.
Visualization of collagen fibers in cardiac tissues is essential for clinical diagnosis and pathological analysis of cardiac fibrosis. Selecting a proper imaging method is still challenging for researchers and clinicians who want to determine specific information about the collagen network in cardiac tissues. We examined fibrillar collagen network from mouse ventricular myocardium by commonly available light microscopy techniques using our home-built multimodal microscope. Myocardial slices were unstained or stained with either Picrosirius red or collagen type I antibody/dye conjugation, then imaged by polarized light, confocal fluorescence, second harmonic generation (SHG), two-photon excited fluorescence (TPEF), and stimulated emission depletion (STED) microscopy techniques. This study is intended to serve as a reference for basic research and clinical evaluation of fibrillar collage network in cardiac tissues.
Understanding cardiomyocyte-extracellular matrix (ECM) interactions at the molecular level is essential for deeper insights into their mechanical signaling function for cardiac development, homeostasis and remodeling. We report a lab-built microscope integrating two-color STED microscopy with second harmonic generation (SHG) microscopy to investigate the detailed architecture of cardiomyocyte-ECM interactions in murine myocardium at a subdiffractive level. SHG microscopy is used to locate possible interaction sites at the cell-ECM interface through the intrinsic SHG signal generated by collagen assemblies and myosin filaments. Two-color STED microscopy is used to obtain a subdiffractive view of proteins at sites of interest registered by SHG microscopy. Because large field-of-view (FOV) STED microscopy is still challenging, with photobleaching often a major concern, imaging only SHG-registered sites is advantageous. Further, using intrinsic contrast in the study reduces the number of biomarkers for fluorescent staining and thereby the number of detection channels for fluorescent imaging, simplifying sample preparation procedures and STED microscopy architectures. For purpose of demonstration, we show images of immunostained type I collagen, type Ⅳ collagen and laminin as ECM structures of interest in rat ventricular sections without counterstaining.
Chondrocyte viability is an important measure to consider when assessing cartilage health. Dye-based cell viability assays are not suitable for in vivo or long-term studies. We have introduced a non-labeling viability assay based on the assessment of high-resolution images of cells and collagen structure using two-photon stimulated autofluorescence and second harmonic generation microscopy. By either the visual or quantitative assessment, we were able to differentiate living from dead chondrocytes in those images. However, both techniques require human participation and have limited throughputs. Throughput can be increased by using methods for automated cell-based image processing. Due to the poor image contrast, traditional image processing methods are ineffective on autofluorescence images produced by nonlinear microscopes. In this work, we examined chondrocyte segmentation and classification using Mask R-CNN, a deep learning approach to implement automated viability analysis. It has been demonstrated an 85% accuracy in chondrocyte viability assessment with proper training. This study demonstrates that automated and highly accurate image analysis is achievable with the use of deep learning methods. This image processing approach can be helpful to other imaging applications in clinical medicine and biological research.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.