This PDF file contains the front matter associated with SPIE Proceedings Volume 11875, including the Title Page, Copyright information, and Table of Contents.

Welcome and Introduction to Conference 11875

As optical systems become smaller and requirements for packaging and functional performance demand more unique solutions to traditional imaging problems, the opportunity for exploration and advancement in simulation and design of non-traditional systems has grown considerably. For example, multi-layer diffractive elements and metalenses play a part in this new world of tiny optical systems, and in this paper we will explore some example system designs showing hybrid approaches from both traditional imaging and diffractive optical design, combined with more rigorous vector electromagnetic wave propagation. We will show combinations of phase optimization and subsequent nano-cell creation in full vector tools, as well as unique propagation of the electromagnetic field in combination from a rigorous model and a simplified beamlet-based decomposition approach. These tools can play a significant role in the design, optimization and analysis of these unique systems both now and in the future.

In recent years, we have seen the development of integrated plenoptic sensors, where multiple pixels are placed under one microlens. It is mainly used by cameras and smartphones to drive the autofocus of the main lens, and it often takes the form of dual-pixels with 2 rectangular sub-pixels. We study the evolution of dual-pixels, the so-called quad-pixel sensor with 2x2 square sub-pixels under the microlens. As it is a simple light field capturing device, we investigate the computational photography abilities of such sensor. We first present our work on pixel-level simulations, then we present a model of image formation taking into account the diffraction by the microlens. Finally, we present new ways to process a quad-pixel images based on deep learning.

We present py_pol, an open source library developed in Python. It can be used to perform polarization calculations both in Jones and Stokes-Mueller formalism, so it can be used from simple to very complex problems. It allows creating and manipulating light states and optical elements, calculating and plotting many parameters and checks, and even performing advanced algorithms to filter experimental errors. Also, it is optimized for managing multiple light states or optical elements simultaneously requiring very few computation resources. It has an extensive documentation, allowing the user to easily learn the use of the library.

Photoacoustic tomography technology is a new non-invasive, non-ionizing biomedical imaging method. This technology combines the high contrast of optical imaging and the high-resolution characteristics of ultrasound imaging, which can obtain high-resolution images in deeper tissues. In recent years, it has developed rapidly and won widespread attention. Traditional sampling method must follow the Nyquist sampling theorem, which wastes a lot of sensing time and storage space. In order to improve the sampling efficiency, compressed sensing (CS) theory is used to collect and process photoacoustic data. The advantage of CS theory is that it can combine data acquisition and data compression. So that only the sparse features of the original signal need to be collected, and a high-quality original target image can be successfully reconstructed with very few samples, which greatly reduces data redundancy. More than that, the requirements for equipment are reduced. This paper uses MATLAB's k-wave simulation toolbox to establish a virtual photoacoustic field, collect the photoacoustic signals of biological tissues, and reconstruct the image through the segmented weak orthogonal matching pursuit (StOMP) algorithm. The results show that the MATLAB virtual compressed sensing photoacoustic tomography simulation platform based on k-wave can realize high-quality photoacoustic tomography with less data. The superiority of the compressed sensing theory and the efficiency of the k-wave virtual platform are verified.

In this work, the problem of designing proper Phase-Shifting Masks (PSMs) suitable for optogenetic applications is considered. In such applications, structured light is used to stimulate neurons or groups of neurons while short-term excitation is required to study the dynamics of the neuronal activity. In practice, such fast response times can be achieved only via the use of ferroelectric Spatial Light Modulators (fSLMs) that posses significantly smaller response times as compared to the, more common, liquid crystal based SLMs. However, typical fSLMs are restricted to using only a small number of discrete phase levels. To this end, we propose a regularized cost function for Phase-Shifting Mask design, that promotes phases in a discrete phase set. Significantly higher Peak Signal-to-Noise Ratio (PSNR) is achieved by the proposed approach, as compared to other approaches.

Photoacoustic microscopy (PAM) is an imaging technology developed rapidly in recent years. The technology has the advantages of high resolution, rich contrast of optical imaging and high penetration depth of acoustic imaging. It is widely used in biomedical field, such as tumor detection. Photoacoustic images can not only reflect the structural characteristics of tissues, but also reflect the metabolic state, disease characteristics and even nerve activity of tissues, so as to realize functional imaging. Photoacoustic (PA) signals are inherently recorded in noisy environments and are also exposed to the noise of system components. The presence of noise has a great negative impact on image quality and interferes with image details. Therefore, it is necessary to reduce the noise in PA signals to reconstruct images with less interference information. Because deep learning can process image information quickly and efficiently, deep learning has become the preferred method for photoacoustic image denoising in recent years. In this study, the photoacoustic blood vessel image obtained was added with a certain intensity of Gaussian noise, and the denoising generative adversarial network based on Wasserstein distance (WGAN) was used to denoise the photoacoustic blood image. For the purpose of evaluation, the Peak Signal-to-Noise Ratio (RSNR), Structural Similarity Index Metric (SSIM), Universal Quality Index (UQI) and Image Enhancement Factor (IEF) were calculated. According to the calculation results, this study effectively improves the image quality, proves the effectiveness of the neural network, and has good clinical significance and broad application prospects.

Light propagation in non-homogeneous media with varying refractive index is challenging to analyse, even with knowledge of the gradient index (GRIN) structure n(x,y,z). Differential equations can be used to describe the path light rays take when travelling through a medium. Solutions to a differential equation are used to calculate the focal length of a GRIN lens with focus on quadratic GRIN profiles. Ray-tracing in GRIN media is notoriously difficult, and usually not possible without numerical methods, however for several types of GRIN lenses, the analytical solution exists. Examples of spherical GRIN lenses can be found in nature such as the lens in an octopus eye. A formula is proposed here to calculate the focal length of such a lens which will be compared to previous studies and models created on ray-tracing software, ZEMAX.

Realistic rendering that relies on physically correct laws of propagation and accumulation of light energy, is used for solving a wide range of applied problems, including virtual prototyping of complex optical systems. With the increasing computation efficiency and complexity of computational architecture, both rendering complexity, and the required computational accuracy increase. Taking into account that modern workstations might have several CPUs with up to 128 virtual cores each, the task of the effective parallelization of the rendering algorithms that utilize all CPU resources is an urgent challenge. In the scope of the current research, the authors investigated the application of various CPU parallelization approaches for the realistic rendering algorithms based on the backward photon mapping, and their limitations. These methods include traditional methods such as synchronous and asynchronous parallelization approaches and their combination. As a result of the research, the authors developed the three-level parallelization method, consisting of fully synchronous, partially synchronous, and asynchronous levels. The key feature of the three-level parallelization method is the additional semi-synchronous level with shared memory. Due to the use of semi-synchronous calculations and asynchronous data exchange between threads, there is no need to synchronize the access to a shared data, which results in increased rendering speed. The three-level parallelization model can also be used in distributed systems due to the asynchronous model is used at the top level. The results of the comparative testing of parallelization approaches on the multiprocessor workstation when rendering images formed by virtual prototypes of the real systems are presented.

A model-based edge detection is always required when quantitatively evaluating the bidirectional measurements of micro- and nanostructures in optical microscopy. For example, the accurate determination of the width of a structure requires the knowledge of the location of the real physical edges in the measured profile. The interpretation of the measured edge profile cannot be performed intuitively due to distortion which is caused by diffraction and refraction. We advise a model-based edge detection algorithm which is based on rigorous simulations of the microscope’s imaging. The intensity level which corresponds to the position of the real physical edge is called the threshold and it is determined in the simulations. For these optical simulations we employ the JCMsuite, which is a software application of the finite-element-method (FEM). Since numerical and semi-analytical methods for the calculation of electromagnetics in optical systems rely to some degree on approximations, their results may vary even when the input parameters are identical. We apply a test suite of input parameters for the purpose of comparing numerical simulation tools regarding the resulting thresholds for measurements on line-shaped nanostructures in a periodic grating. The test suite maintains the illumination and imaging parameters of a transmitted light UV-microscope while the object parameters of a binary line grating are varied. There are 25 grating configurations with different line-to-space ratios, where the line width ranges from the resolution limit up to almost 10 µm. The illumination pupil is discretized in a cartesian grid with 113 grid points in total. We introduce different pupil samplings, after calculating the threshold values of the original test suite. We obtain a high agreement of the thresholds results and the related linewidth values when comparing with already performed results of two additional rigorous applications. Furthermore, we showcase the threshold variation for different samplings of the illumination pupil.

High-power laser beams are usually imaged with the help of suitable optics, consisting of several optical elements. The propagation of such high-power beams can result in a significant thermally induced focus shift, the thermal lens effect. This is mainly due to the deposited power of the beam itself, acting as a heat source in the optical elements. The resulting change in temperature leads to both, a deformation of the optical element, and a change of the refractive index. In this work, the finite element method was coupled with ray tracing, in order to get both, optical and thermal-structural simulation results. Additionally, thermal radiation and the mutual influence of optical elements have to be considered. With the method proposed, the thermal lens of optical systems can be determined exactly, but it also enables for designing of new methods to completely compensate for thermal lensing. In this work, a complete compensation of thermal lensing for a F-θ-scanner could be simulated, which even works at optical powers of 10kW.

Microscopic imaging of anisotropic samples has many important applications in cytopathology. The endogenous contrast from the polarization properties of a specimen, such as its birefringence, provides valuable diagnostic information for several deadly diseases, including cardiac amyloidosis and squamous cell carcinoma, for example. In the past, polarized light microscopy (PLM) has been widely used as a diagnostic tool during the clinical review. However, in analogy with the standard microscope, the PLM typically has a restricted spatial-bandwidth product (SBP). As a consequence, one can either image a large area with low resolution or see the details of a very small area of the sample at the resolutions required for accurate analysis. To address the SBP issue of the PLM, we propose a computational microscopy method, termed vectorial Fourier ptychography, to illuminate the specimen with polarized light from different angles and detects different polarization states of the diffracted light. By illuminating a specimen with plane waves from different angles, our vectorial Fourier ptychography method effectively modulates the high-spatial-frequency components of the specimen into lower frequencies that can be detected by the optical system. With a Jones calculus-based forward model and a second-order phase retrieval method, we can reconstruct high-resolution, wide field-of-view(FOV) amplitude, phase, birefringence, retardance, and diattenuation of the specimen. To assess the reconstruction accuracy of our method, we imaged polystyrene beads submerged in immersion oils of different refractive index, as well as monosodium urate crystals. Further, To validate the diattenuation reconstruction accuracy, we reconstruct a USAF resolution test chart with a half blocked by a linear polarizer. These experiments confirm quantitatively accurate reconstruction results with a 1.25 um full-pitch resolution over a FOV of 6.6 x 4.4 mm^2, which is 5 times higher than the native (brightfield) resolution of the non-computational optical system. Finally, we demonstrate our technique by producing high SBP polarization images of several anisotropic biologic samples, includes collagen tissue, congo red stained cardiac tissue, and a bean root sample.

As laser diodes (LDs) replace LEDs in the remote phosphor setup, a new class of lighting solutions emerges, giving rise to laser-excited remote phosphor (LERP) systems. While already in use in some commercial applications such as automotive lighting, these systems have not yet matured. The optical behavior of phosphors is temperature dependent, specifically the absorption coefficient, the conversion efficiency reflected in the quantum efficiency (QE) coefficient, and, to a lesser extent, the emission spectrum. For this reason, opto-thermal analysis is critical for further investigating and optimizing these systems. A steady-state opto-thermal simulation scheme that combines ray tracing in OpticStudio software with heat transfer calculations using the finite element method (F.E.M.) in ANSYS is presented and experimentally validated here. Furthermore, the temperature-dependent models established for phosphor properties are used to optimize the phosphor sample.

In recent years, more and more organizations and teams in the world are engaged in photoacoustic imaging research. Photoacoustic imaging is in a state of vigorous development. As an important branch of photoacoustic microscopy, optical resolution photoacoustic microscopy combines the advantages of optical imaging and acoustic imaging, which has the advantages of high resolution, high contrast, high sensitivity and non-invasiveness. However, in order to obtain high resolution, it is often necessary to focus the laser beam, which will lead to small imaging depth of field and unable to obtain large-scale structural information. However, in clinical diagnosis, doctors want to obtain large-scale and high-resolution structural and functional information as much as possible, so it is of great significance to solve the problem of small depth of field in photoacoustic microscopy. In order to expand the depth of field of photoacoustic microscopy imaging, this paper proposes a three-dimensional information fusion algorithm for photoacoustic microscopy imaging. Firstly, we obtain two sets of vascular data (except the focus position) by virtual photoacoustic microscopy. Then we take out the B scan data of two sets of three-dimensional data sets in turn, and use the fusion algorithm based on pyramid transform to fuse them. Finally, the maximum projection is used to restore the original data and the fused data. We compare the maximum projection before and after fusion. The experimental results show that the algorithm realizes the extension of the depth of field, and the fused data successfully displays more abundant vascular information in an image, and maintains the advantages of high contrast and high resolution of photoacoustic microscopy imaging.

We implement a data efficient approach to train a conditional generative adversarial network (cGAN) to predict 3D mask model aerial images, which involves providing the cGAN with approximated 2D mask model images as inputs and 3D mask model images as outputs. This approach takes advantage of the similarity between the images obtained from both computation models and the computational efficiency of the 2D mask model simulations, which allows the network to train on a reduced amount of training data compared to approaches previously implemented to accurately predict the 3D mask model images. We further demonstrate that the proposed method provides an accuracy improvement over training the network with the mask pattern layouts as inputs. Previous studies have shown that such cGAN architecture is proficient for generalized and complex image-to-image translation tasks. In this work, we demonstrate that adjustments to the weighing of the generator and discriminator losses can significantly improve the accuracy of the network from a lithographic standpoint Our initial tests indicate that only training the generator part of the cGAN can be beneficial to the accuracy while further reducing computational overhead. The accuracy of the network-generated 3D mask model images is demonstrated with low errors of typical lithographic process metrics, such as the critical dimensions and local contrast. The networks predictions also yield substantially reduced the errors compared to the 2D mask model while being on the same level of low computational demands.

High-resolution imaging at short wavelengths from extreme ultraviolet to hard X-rays has many applications in a plethora of fields from astronomy to biology and semiconductor metrology. Unfortunately, efficient optics for these wavelengths are difficult to manufacture or have limited resolution. For this reason, in the past few years, coherent diffraction imaging (CDI) applications become widely used. In CDI, the object is illuminated by a coherent beam and the diffraction intensity is collected by a 2D pixel detector. In this process, the phase information of the diffracted light is lost. A phase retrieval algorithm is then used to reconstruct the object’s complex amplitude. Ptychography is a scanning version of coherent diffraction imaging and it is based on an iterative reconstruction algorithm that relies on the quality of the recorded diffraction intensity to converge. To obtain diffraction patterns with a high signal-to-noise ratio, a beam stop is used in many ptychography setups to avoid over-saturation and blooming effects on the detector. While using a beam stop in a ptychography setup has become common practice, the limits of affordable data loss due to beam stop have not been systematically investigated. Pixel masking is the conventional method to recover the lost frequencies. In this method, when enforcing the Fourier domain constraint, the invalid pixels are ignored. In the missing data region, the algorithm is allowed to keep the guess from the previous iteration. The illumination conditions of the ptychography experiment play a critical role in the signal recovery procedure. The diffraction pattern on the detector is the convolution of the Fourier transform of the object and the illumination. An illumination with a finite numerical aperture encodes the object information over a larger detector area. This makes the reconstruction algorithm more robust to pixel loss. We provide simulation and experimental results to demonstrate this theory.

This paper discusses a collaborative effort of two Fraunhofer institutes to develop a lithography model that simulates the fabrication of blazed gratings using grayscale lithography. The model is calibrated with experimental data of blazed grating profiles. The complete process of modeling and calibration has been performed using the research and development lithography simulator Dr.LiTHO. To emulate the grayscale exposure of blazed gratings in a LED-based micro-image stepper with Dr.LiTHO a thin mask with a linear variation of the mask transmission and corresponding distribution of exposure dose was used. The resulting photoresist profiles are obtained with a standard model for Diazonaphthoquinone (DNQ) photoresists. The calibration of simulated and experimental profile data of blazed gratings is performed using Dr.LiTHO’s inbuilt optimizer - Pythmea. The difference between experimental and simulated profile shapes is expressed by an areaFit. Minimization of this areaFit versus photoresist parameters and correlation analysis help to identify the most appropriate model parameters.

The impact of polarization was observed in the extreme ultraviolet (EUV) imaging simulations for high NA lithography [3] [4] [5]. It is shown that polarized illumination can improve the local contrast of images or NILS (normalized intensity log slope). This work investigates the possibilities to polarize EUV light by optimized multilayers. The characterization and simulation of multilayer structures has been performed using Dr.LiTHO [10]. The most efficient multilayer polarizers operate close to Brewster angle, where the reflectivity for TM polarized light (RTM) is close to zero, according to Fresnel’s equations. A multiobjective optimization algorithm was used to identify the suitable multilayer configurations maximizing reflectivity of TE polarized light (RTE) and fraction of polarization. Fraction of polarization (FoP) was calculated as the ratio between (RTE-RTM)/(RTE+RTM) to obtain the suitable multilayer with variable thickness. The multilayer structure is optimized to have the highest reflectivity of TE polarized light and fraction of polarization at the Brewster angle. It was found that MoSi multilayer can achieve 99.9% fraction of polarization by optimizing the thickness of Si and Mo. In reality, a multilayer polarizer has to operate over certain ranges of incident angles and/or wavelength ranges. Multilayer is optimized for different ranges of wavelength (13 nm : 14 nm) and incidence angles (37° : 47°). Additional simulations investigate the impact of different options in the design of the multilayer (e.g., constant vs. variable bilayer thickness) and materials (e.g., RuSi vs. MoSi multilayers) on the achievable performance.

EUV photomasks define the lithographic layers of chips, which are binary structures of relatively low versatility in contrast to specimen in biology or materials science. This knowledge can be used in EUV photomask imaging and inspection methods to improve the speed or sensitivity. We present here a total variation-based phase retrieval algorithm similar to previous methods by Chang et al. and Enfedaque et al. for EUV mask imaging and metrology. Total variation (TV) regularization exploits the binary structure of the reticles, enforcing a sparse sample gradient. We compare the TV regularized algorithm, PtychoADMM, to a standard phase retrieval approach, the difference map (DM). For simulated data containing Poisson noise, we do not observe a benefit from using the TV based PtychoADMM algorithm. The reconstructed image quality is similar, while PtychoADMM being a computationally more demanding method. In future, we will investigate if TV can recover information where the standard DM approach fails, e.g. for relaxed measurement requirements like a lower signal to noise ratio or less probe overlap in the ptychography scan.

As a new non-destructive medical imaging technology, photoacoustic imaging combines the advantages of optical imaging and ultrasonic imaging, which has the characteristics of high contrast and strong penetration ability. It can effectively image biological tissues and functions, and be applied to the early diagnosis and treatment of tumors, cardiovascular and cerebrovascular diseases, which has broad prospects for development in the field of biomedicine. When photoacoustic imaging is used to collect a large number of pathological medical images, it is prone to slow data transmission and poor reconstruction effect. In an effort to speed up data transmission and improve image reconstruction quality, this article is based on photoacoustic imaging technology and compressed sensing reconstruction algorithm, a virtual simulation platform of photoacoustic tomography combined with compressive sensing is built by using K-wave simulation toolbox, and the hard thresholding pursuit algorithm is used to complete the signal reconstruction. In order to verify the performance of the virtual simulation platform, in this paper, the local vascular network is compressed and reconstructed. The reconstructed image retains the main information of the original image, and the edge features are similar. The results show that the virtual simulation platform can reconstruct high quality images by a small amount of data, which provides important significance and theoretical research value for the application of the compressed sensing in photoacoustic imaging.

The development and research of new devices and systems of diffractive and integrated optics, which based on elements with thin-layer micro- and nanostructures, requires the improvement of the technological base. The most massive and lowcost is planar optical elements, on the surface of which diffractive and raster computer-synthesized micro- and nanostructures are formed, as well as structures based on synthesized metamaterials. In recent years, off-axis and axisymmetric computer-synthesized holograms for control and alignment of optical systems, microstructured optical elements with 3D microrelief for complex transformations of wavefronts and intensity distributions of light beams (micro-optics), integrated-optical passive and active circuits have been greatly developed. In this paper describes the principles of operation of two different scanning laser nanolithography system developed operating in a writing and polar coordinate system. Development and research work these lithographic systems were conducted at the Institute of Automation and Electrometry of the SB RAS for many years. The areas of applicability of these systems are described, their differences and technical limitations are considered. The emphasis is made on fundamentally similar units of installations of this class and the prerequisites for their unification are considered. Methods for increasing their resolution, speed and accuracy of writing are proposed, prospects and directions of their development are analyzed. The results of writing test optical elements on metal films using the described methods are demonstrated.

As an advanced technology, polarization imaging has attracted widespread interests in recent years due to its unique ability to detect the polarization information of objects. With illumination of partially coherent and partially polarized light, the polarization imaging systems are in generally nonlinear systems because they do not possess a transfer matrix in the usual sense. Nonetheless, it is possible to apply a sinusoidal amplitude object at the input and to measure a periodic output. However, unlike the linear case, knowledge of the response of the polarization system to a sinusoidal input does not allow one to predict what the output will be for other types of inputs. To address the fundamental nonlinearity in evaluation of the Stokes images, we provide a frequency- domain calculation of Stoke images of a sinusoidal amplitude grating under partially polarized and partially coherent illumination, and propose a new concept referred to as transmission cross coefficients to describe its frequency response of polarization imaging system. Some of the implications of such analysis are also given for polarization imaging evaluation.

In this paper, we give a spatial frequency analysis of polarization imaging system and study the effects of rotational alignment errors of the polarizer from their ideal orientations. With the help of newly proposed Optical Transfer Matrix (OTM) for a diffraction-limited polarization imaging system, we investigate the effects of polarization-sensitive aberrations from the generalized pupil matrix at the exit pupil plane. Here, polarization aberrations stemming from the angular alignment errors of a linear polarizer has been demonstrated. The performance of a polarization imaging system with rotational alignment error has also been evaluated based on a cost function based on OTM.

In this paper, we revisit the role of polarization and coherence in the study of polarization speckle. It is shown that the type of polarization and coherence used in such analysis has an important difference from the "classical" polarization and coherence concepts. The analogy between polarization speckle and partially polarized thermal radiation is explored. We propose a concept referred to as ensemble-average polarization and coherence for statistical optics and give the definition and physical indication for polarization speckle. Some statistics associated with polarization speckle including the 1st order statistical statistics of the Stokes parameters and the ensemble-average Van Cittert-Zernike theorem for the ensembled-average generalized Stokes parameters are investigated theoretically and experimentally to demonstrate their link and different physical features as compared with the conventional concepts.

Photothermal imaging plays an important role in brain structural and functional imaging. However, the skull has a strong scattering effect on photons, it is necessary to study the propagation behavior and thermal effect of photons in brain. In this study, MCmatlab, an open source program in MATLAB, which combines the Monte Carlo method with finite element method, was used to build a photothermal model of mouse brain to simulate the movement of photons in mouse brain. Monte Carlo method is commonly used to solve the distribution of light in biological tissues, and FEM (Finite Element Method) can effectively solve the distribution of optical parameters in biological tissues, and helps to obtain an approximate solution of the radiation transfer equation (RTE). By constructing a 3D model to observe the structure of each layer of the brain more intuitively, the laser pulse propagation in the mouse brain and the temperature distribution of each layer of brain tissue were obtained by the RTE solver and the finite element thermal diffusion solver included in MCmatlab.

Publisher’s Note: This paper, originally published on 13 September 2021, was replaced with a corrected/revised version on 27 September 2021. If you downloaded the original PDF but are unable to access the revision, please contact SPIE Digital Library Customer Service for assistance.