Fast denoising and lossless spectrum extraction in stimulated Raman scattering microscopy

Deep learning-based high-speed, large-field, and high-resolution multiphoton imaging

Multiphoton microscopy is a formidable tool for the pathological analysis of tumors. The physical limitations of imaging systems and the low efficiencies inherent in nonlinear processes have prevented the simultaneous achievement of high imaging speed and high resolution. We demonstrate a self-alignment dual-attention-guided residual-in-residual generative adversarial network trained with various multiphoton images. The network enhances image contrast and spatial resolution, suppresses noise, and scanning fringe artifacts, and eliminates the mutual exclusion between field of view, image quality, and imaging speed. The network may be integrated into commercial microscopes for large-scale, high-resolution, and low photobleaching studies of tumor environments.

Multidimensional quantitative characterization of the tumor microenvironment by multicontrast nonlinear microscopy

Characterization of the microenvironment features of tumors, such as its microstructures, biomolecular metabolism, and functional dynamics, may provide essential pathologic information about the tumor, tumor margin, and adjacent normal tissue for early and intraoperative diagnosis. However, it can be particularly challenging to obtain faithful and comprehensive pathological information simultaneously from unperturbed tissues due to the complexity of the microenvironment in organisms. Super-multiplex nonlinear optical imaging system emerged and matured as an attractive tool for acquisition and elucidation of the nonlinear properties correlated with tumor microenvironment. Here, we introduced a nonlinear effects-based multidimensional optical imaging platform and methodology to simultaneously and efficiently capture contrasting and complementary nonlinear optical signatures of freshly excised human skin tissues. The qualitative and quantitative analysis of autofluorescence (FAD), collagen fiber, and intracellular components (lipids and proteins) illustrated the differences about morphological changes and biomolecular metabolic processes of the epidermis and dermis in different skin carcinogenic types. Interpretation of multi-parameter stain-free histological findings complements conventional H&E-stained slides for investigating basal cell carcinoma and pigmented nevus, validates the platform's versatility and efficiency for classifying subtypes of skin carcinoma, and provides the potential to translate endogenous molecule into biomarker for assisting in rapid cancer screening and diagnosis.

Noises investigations and image denoising in femtosecond stimulated Raman scattering microscopy

In the literature of SRS microscopy, the hardware characterization usually remains separate from the image processing. In this article, we consider both these aspects and statistical properties analysis of image noise, which plays the vital role of joining links between them. Firstly, we perform hardware characterization by systematic measurements of noise sources, demonstrating that our in-house built microscope is shot noise limited. Secondly, we analyze the statistical properties of the overall image noise, and we prove that the noise distribution can be dependent on image direction, whose origin is the use of a lock-in time constant longer than pixel dwell time. Finally, we compare the performances of two widespread general algorithms, that is, singular value decomposition and discrete wavelet transform, with a method, that is, singular spectrum analysis (SSA), which has been adapted for stimulated Raman scattering images. In order to validate our algorithms, in our investigations lipids droplets have been used and we demonstrate that the adapted SSA method provides an improvement in image denoising.

Super-Multiplex Nonlinear Optical Imaging Unscrambles the Statistical Complexity of Cancer Subtypes and Tumor Microenvironment

Label-free nonlinear optical imaging (NLOI) has made tremendous inroads toward unscrambling the microcosmic complexity of cancers. However, harmonic and Raman microscopy offers throughput without redox information to reveal metabolic differentiation, and fluorescence lifetime microscopy lacks the vibrational response of molecules to visualize specific molecular constituents such as lipid. Here, a flexible, robust simultaneous multi-nonlinear imaging and cross-modality system that combines complementary imaging contrast mechanisms is demonstrated. This system, utilizing multiplexed ultrashort pulses, ingeniously integrates typical nonlinear processes, and high-dimension lifetime extension in a single setup to enhance the imaging dimensions and quality. Using this system, the authors perform label-free comprehensive evaluation of clinicopathological tissues of ovarian carcinoma due to its statistical complexity. The results show that the technology provides statistically rich, insightful information with high accuracy, sensitivity, and specificity, in contrast to standard histopathology, and can potentially be a powerful tool for fundamental cancer research and clinical applications.