This PDF file contains the front matter associated with SPIE Proceedings Volume 12601, including the Title Page, Copyright information, Table of Contents, and Conference Committee lists.

Deep learning is rapidly becoming the state of the art, leading to enhanced performance in various medical diagnosis. In recent years, convolutional neural network (CNN) has been used to diagnose retinal disease and has proven its superiority in detection and classification tasks. Vision transformer is a new image classification model that has been proposed in 2020. It does not rely on any CNN and completely performs based on the transformer structure which has a different feature extraction method from CNN. In this study, diagnosis of retinal disease using vision transformer was presented using optical coherence tomography (OCT) images. A multi-class classification layer in the vision transformer model was used to group the OCT images into the normal and three abnormal type, Choroidal Neovascularization (CNV), Drusen, and Diabetic Macular Edema (DME). The proposed method achieved a accuracy of 95.76%, sensitivity of 95.77% and specificity of 98.59% in detecting CNV, DME and DRUSEN. Results showed that the classification accuracy of vision transformer is higher than that of other traditional CNN models. The performance of vision transformer was evaluated with different performance metrics like accuracy, sensitivity, and specificity, which proved that vision transformer is a statistically significant method than other standard CNN architectures in classifying retinal diseases using OCT images. This technology enables early diagnosis of retinal diseases, which may be useful for optimal treatment to reduce vision loss.

Hyperspectral microscopic imaging (HMI) technology is a non-contact optical diagnostic method, which combines hyperspectral imaging (HSI) technology with microscopy to provide both spectral information and image information of the samples to be measured. In this paper, basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and malignant melanoma (MM) were classified based on synthetic RGB image data from HMI cube by using four classification methods extreme learning machine (ELM), support vector machine (SVM), decision tree and random forest (RF). The highest classification accuracy of 0.791±0.060 and a KAPPA value of 0.685±0.095 were obtained when color moment, gray level co-occurrence matrix (GLCM) and local binary pattern (LBP) were used for image feature extraction, feature dimensions were reduced by the PLS, the sample sets were divided by the hold-out method, and the tissues were classified by the SVM model.

As an important branch of photoacoustic microscopy, optical-resolution photoacoustic microscopy suffers from limited depth of field due to the strongly focused laser beam. In this work, a 3D information fusion algorithm based on 3D stationary wavelet transform and joint weighted evaluation optimization is proposed to fuse multi-focus photoacoustic data to achieve large-volumetric and high-resolution 3D imaging. First, a three-dimensional stationary wavelet transform was performed on the multi-focus data to obtain eight wavelet coefficients. Differential evolution algorithm based on joint weighted evaluation was then employed to optimize the block size of division for each wavelet coefficient. Corresponding sub-coefficients of multi-focus 3D data were fused with the proposed fusion rule utilizing standard deviation for focus detection. Finally, photoacoustic microscopy with large depth of field can be achieved by applying the inverse stationary wavelet transform on the 8 fused sub-coefficients. The fusion result of multi-focus vertically tilted fiber shows that the depth of field of optical-resolution photoacoustic microscopy is doubled without sacrificing lateral resolution via the proposed method. The effectiveness of the proposed method was verified through the fusion results of multi-focus vessel data.

Ovarian cancer is a disease with a high mortality rate in women. The important reasons for high mortality rate of ovarian cancer is the difficulty in early detection. The process of cell carcinogenesis is often accompanied by changes in surface nanostructure of cell membrane. In this study, atomic force microscopy (AFM) was used to obtain the nanostructure features of ovarian cancer cells. IOSE-80 (human ovarian normal cells) and Caov3 (human ovarian cancer cells) cell lines were selected and the morphology of the cell nuclear regions were measured using AFM Quantitative Imaging (QI) mode, which can offer information of hight, adhesion and slope channels. The surface parameters of the cell obtained from the three channels were analyzed. The results showed that there were significant statistical differences in parameters Root-mean-square height (Sq), Skewness (Ssk), Maximum height (Sz) and Arithemetic mean height (Sa) of adhesion channel, Sq, Ssk and Sa of hight channel. These findings indicate that the three channel in AFM imaging can offer different information of the surface nanostructure and the combination of these feature parameters may improve the identification accuracy of cancer. Our study will provide a new idea for the early diagnosis of ovarian cancer based on the nanostructure features of cell surface at the single-cell level.

Prostate cancer is the 2nd most commonly occurring male cancer and the 4th most common cancer overall. Early detection and diagnosis are important for clinical treatment. Atomic force microscopy (AFM)-based techniques have been shown to have potential in detecting malignant cancers and artificial intelligence can improve the accuracy of diagnostic and prognostic prediction tests. In this study, the classification of AFM images of prostate cells was performed using machine learning. For early prediction, we used the support vector machine (SVM) to classification prostate cells and compare the classification performance with the remaining four conventional classifiers such as logistic regression (LR), stochastic gradient descent (SGD), K-nearest neighbours (KNN), random forest (RF). Most of the classifiers did well after using the feature selection method (BorutaShap). The results show that the accuracy (ACC) of the features selected using the BorutaShap algorithm combined with the SVM classifier can reach 82.5%. Our current study demonstrates that AFM imaging combined with machine learning can be used to identify prostate cancer cells with an effective classification performance and robustness.

Piezoelectric ceramic tube fiber-scanning two-photon endomicroscopy is an essential division of miniature two-photon microscopy. The reverse collection optical path of the two-photon endomicroscopy platform is modeled and designed in this study. After simulating the chromatic aberration characteristics of the objective, the effects of the collection signal wavelengths, off-axis positions, fiber cladding diameters, and imaging depths on the collection efficiency are evaluated using Monte Carlo simulation. The results provide an additional theoretical explanation for enhancing the two-photon endomicroscopy platform's imaging sensitivity and signal-to-noise ratio.

In this work, a multiband microwave waveforms transmitting system was demonstrated based on all-optical dual fiber combs. Two optical fiber frequency combs via cascaded four waves mixing process ensure the coherence of the system by sharing the pump light. The high-speed frequency tuning of the pump laser will be transferred to each newly generated comb line through the nonlinear effect in the fiber. The comb pairs of the same order obtained through filtering transform the frequency chirp in the optical domain to the frequency chirp in the electrical domain after being beat on the PD. By changing the parameters of the control signal, the center frequency, bandwidth, PRF and envelope of the multi band microwave signal can be flexibly reconfigurable. Thanks to the coherence of dual optical frequency combs, microwave signals all show excellent line-type, high signal to noise ratio and no burrs. As a result, the chirped microwave signals of triangle wave, sine wave and sawtooth wave types are obtained at the repetition rate of 500 Hz. Overall, multi-band chirped microwave generation system based on all-optical system provides a practical device for transmitting and receiving multi band radar.

The four-optical coherent mixing detection technology can improve the dynamic range of moving target detection. This method has the difficulty of distinguishing the type of mixing output signal. We propose a method to distinguish the signal type by using the different peaks of the mixed signal spectrum. Based on the statistical theory, the power spectrum function of the mixed signal is obtained, and the numerical analysis of the influence of the light source line width and the light source frequency difference on the signal power spectrum is carried out. Through numerical calculation and analysis, the results show that the increase of the light source linewidth will lead to the broadening of the signal power spectrum. When the Doppler frequency difference is greater than 1/5 times the linewidth of the light source, the power spectrum of the two homodyne coherent signals in the four-light coherent mixing can be distinguished; when the Doppler frequency difference is less than 1/5 times the light source linewidth , The power spectrum of the two homodyne coherent signals in four-optical coherent mixing can not be distinguished.

We present the design, fabrication, and performance of high-speed oxide-confined 1030 nm vertical-cavity surfaceemitting lasers (VCSELs) with a short optical cavity, and multiple oxide apertures. High-speed modulation was facilitated by using the shortest possible cavity of a half-wavelength length and multiple oxide apertures to enhance the confinement of optical fields and reduce capacitance. We carefully optimized the multiple quantum wells and the doping profile in order to achieve a high-speed operation. The developed VCSELs exhibit a modulation bandwidth exceeding 25.1 GHz at 25°C, supporting back-to-back data rate up to 40 Gb/s under binary non-return-to-zero (NRZ) modulation.

With the advantages of high communication rate and strong anti-interference, wireless laser communication is very suitable for space BPV communication, UAV directional communication and other scenarios with large-volume data and high reliability requirements. In order to break the traditional one-to-one laser communication scenario, a multi-user modulating retro-reflector(MRR) laser communication system is proposed in this paper, while ensuring the miniaturization of the receiver equipment and realizing multi-user cascade communication. The closed form solution of the outage probability of the proposed multi-user MRR laser communication system under atmospheric turbulence fading is derived, and the influence of system parameters on the system performance is analyzed. The correctness of the derived results is verified by simulation results.

In this paper, we adopt a double-delayed parallel reservoir computing scheme to identify the noise-interfered distorted optical signals with IQ modulation formats (QPSK, 4QAM, 8PSK). Influences of different input layer masks and signal-noise-ratio (SNR) are investigated in detail. Results demonstrate that the proposed photonic reservoir computing can achieve over 97% identification accuracy for distorted QPSK-4QAM or QPSK-8PSK signal sequences.

Optical chains have received widespread attention due to their unique characteristics of high intensity and multiple potential wells. Many methods have been used to generate optical chains, such as using diffractive optical elements or 4Pi systems to modulate vector beams. But these methods require additional phase elements or more complex optical systems. In this work, a single-layer polarization-insensitive metalens with phase distribution of the binary optical element was used to focus the radially polarized beam to generate optical chains. The Richards-Wolf vector diffraction theory was employed to calculate the focal field distribution of the radially polarized beam. Optical chains generated by both the simulated and theoretical calculation are composed of alternating solid points and bubbles which indicates that the theoretical results are in agreement with the simulation results. The introduction of metalens reduces the volume of the optical system, which is conducive to the miniaturization and integration of the optical system. This work may contribute to particle trapping and manipulation, optical micro-nano processing, etc.

Tractor beams have received increasing attention. The generation of tractor beams is an important issue in practical applications. In this paper, Pancharatnam-Berry (PB) metasurfaces were designed to generate non-paraxial Bessel tractor beams. The optical pulling forces (OPFs) exerted on dielectric particles with specific radii on the axis of the non-paraxial Bessel beam were obtained. The presence of OPFs depended on the size of the particles, indicating the potential of the non-paraxial Bessel tractor beam for separating particles. Such a feature illustrated the possibility of selective optical manipulation and sorting. This setup of metasurface has good potential in such aspects as lab-on-a-chip.

Optical tunable filters play a key role in silicon photonic integrated circuits. Highly energy-efficient tunability and a wide continuous tuning range are strongly desired for silicon photonics filters. All-optically thermo-optic (TO) tunable devices based on the light absorbers integrated close to the silicon structure as localized heaters have attracted increasing attention because optical heaters, compared with electrical ones, can greatly reduce thermal loads and heat leakage for the device. They provide a new approach to implementing high-efficiency TO tuning with a fast response. In this work, we propose and experimentally demonstrate an on-chip all-optically tunable filter based on a suspended silicon microdisk resonator with an ultra-compact optical heater, which is a platinum absorber deposited directly on the top of the ridge waveguide. Attributed to the novel optical pumping scheme, ultra-small device size, and suspended waveguide structure, an ultra-high tuning efficiency of 37.70 nm/mW is achieved. Only 1.405 mW pump power is required to tune the single-resonance filter over a wide spectral range of ∼54.5 nm. The demonstrated tunable optical filter has the advantages of high tuning efficiency, compact footprint, and simple fabrication processes, which has significant applications for on-chip all-optical systems.

The demand for high capacity and integration in modern optical communication technology is becoming prominent. Orbital angular momentum (OAM) plays an important role in optical communication. However, there are still challenges to further expand the flexibility and capacity of optical communication in the axial direction. Here, we propose a single-layer liquid crystal device (LCD) to realize the generation of optical vortex (OV) array with arbitrary topological charge in axial multiplane, which can be applied in optical communications based on highly integrated device. The phase of the target OV array is weighted and superimposed to obtain the phase distribution of LCD. In order to obtain an OV array with uniform intensity, it is necessary to determine the optimal weight factor for each OV based on the introduced particle swarm optimization (PSO) algorithm. In the experiment, a LCD with an effective aperture of 2 mm was processed. A CCD captures the OV array image, including two OV arrays at 200*λ (156μm) in front of and behind the focal point respectively. Then, the beam passes through the 4f system of the spatial light modulator with the phase distribution of the Damman vortex grating on the spectrum plane, and the topological charge of the two OV arrays can be detected by the CCD. Our results provide an approach that based on a single liquid crystal plate, OV arrays in multiple propagation planes are realized, in which the number of propagation planes and the position of each propagation plane relative to the back focal plane can be adjusted arbitrarily, and the number, order, and position of OVs in each propagation plane can also be adjusted arbitrarily to meet the application requirements in the field of short-distance optical communication.

The traditional digital differential orthorectification uses a pixel-by-pixel correction method, which is time-consuming and laborious to operate on a single image element and fails to meet the demand for time-sensitive and near real-time DOM production generation. In order to solve this problem, a fast DOM generation method based on multi-pixel binning is investigated in this paper. The method breaks the shackles of the traditional digital differential correction, which only operates on individual pixels, and instead corrects blocks of N×N size, greatly improving the speed of DOM generation processing. The dataset used in this paper are aerial imagery and a DSM of LiDAR acquisition located in Denver, Colorado, USA, to verify the speed and accuracy of the algorithm proposed. Compared with the traditional digital differential correction method, the speed of DOM generation using this method is dependent on the pixel chunk size. The orthorectification speed about 4.1 times faster than the traditional method when using a 3×3 pixels chunks, and about 9.8 times faster than the traditional method when using a 7×7 pixels chunks

Aiming at this problem of the inertial measurement unit has high noise, low precision and large error in traditional attitude calculation methods, an Extended Adaptive Kalman Filter algorithm was proposed to optimize attitude data. The algorithm first builds a state equation model based on sensors such as gyroscope, accelerometer, and magnetometer, with gyroscope data as prediction data, accelerometer and magnetometer measurement values as observation data, and performs error compensation and filtering on the collected raw data. The Seagull Optimization algorithm (SOA) is used to optimize the process noise covariance and measurement noise covariance of the Extended Kalman Filter. Finally, a high-precision aircraft attitude estimation is obtained after Adaptive Extended Kalman Filter algorithm (AEKF) filtering. Both static and dynamic experiments are carried out on the flight experiment platform based on INS-DH-OEM inertial navigation system. Comparing and analyzing the filter effect of the traditional extended Kalman filter algorithm and the adaptive extended Kalman filter algorithm proposed in the paper. Through the experimental results, the algorithm proposed in this paper can suppress the drift of attitude angle, filter out noise and accurately track the attitude change. In the static test, the accuracy of the three attitude angles can be controlled within 0.1°. Compared with the traditional EKF algorithm, the proposed algorithm has better stability and higher accuracy. In the dynamic experiment, the gyroscope has good dynamic performance but with the passage of time, it will produce integral drift, and in this paper, the accelerometer is used to carry out a real-time drift correction of the gyroscope. The roll Angle error and pitch Angle error of this algorithm are within 0.5°, which is higher than that of the traditional EKF algorithm.

A feature point extraction and matching algorithm based on affine transformation space is proposed to address the shortcomings of existing feature extraction and matching algorithms in large view scenes with few effective matching points and slow matching speed. The algorithm first constructs the affine change space to simulate the viewpoint change and obtains the affine invariance; then avoids the feature point detection in the invalid region by dividing the valid region; in the feature description stage, the ORB algorithm is incorporated into the affine change space, while the gradient contrast information of multiple directions in the feature point sampling region is fused to obtain the final binary descriptor. Through experiments on large-view datasets and sequence images, it is demonstrated that the algorithm has better matching effect in large-view scenes, and also has more advantages in time efficiency.

A scheme for measuring a small variation rate of the refractive index (RI) based on orbital angular momentum (OAM) interferometry and time-frequency analysis is proposed and demonstrated in this paper. Two vortex beams carrying OAM of opposite signs are used for interference to produce a petal-like intensity distribution. The variation in RI of the sample leads to a time variable phase delay between the reference and measurement paths, and causes the rotation of the petal-like spots. The rotation angular velocity of the petal-like spots is proportional to the RI variation rate, the normalized cross correlation method is used to estimate it. Then, a time-frequency analysis method is employed to study the time evolution of the variation rate of RI. Three kinds of RI models with different variation rates are simulated and the results are consistent with expectations. The proposed measurement method is simple in structure, and extends a new approach to detect other physical coefficients of RI or the tiny velocity.

In the imaging of low-orbit moving objects, the number of detector elements in the traditional sheared-beam imaging (SBI) system is too great, which seriously restrict the application of SBI. In this paper, the detector array is sparse in two dimensions. We propose a two-dimensional sparse sampling imaging method, which emits a two-dimensional coherent laser array, carries more spectral information of the target at a time and receives speckle echo signals by a two-dimensional sparse detector array for computational imaging. This method can reduce the number of detector elements many times. Firstly, the principle of two-dimensional sparse sampling with SBI detector array is deduced theoretically. Secondly, a two-dimensional spatial sparse reconstruction algorithm is investigated. The target amplitude product and phase difference carried by each detector array element is estimated using discrete Fourier transform, then the target amplitude product and phase difference of all detector array elements are matched respectively to form a complete target amplitude product surface and phase difference surface. The formulas of phase recovery and amplitude demodulation are derived. Finally, the validity and feasibility of the proposed method are verified by simulation. Compared with the traditional three-beam method, when the number of lasers in emission array is M×N, the number of detector elements is reduced to 1/(M-1)/(N-1) of the original without loss of imaging resolution.

Hydrogen energy has quietly become the key to dealing with various environmental problems and energy transformation. With the development of the hydrogen energy industry, safety problems caused by hydrogen leakage also occur continuously. It is particularly important to find a safe and reliable visualization technology to detect hydrogen leakage. Some issues about background oriented schlieren and laser beam profile deformation in visualizing and diagnosing flow fields are summed up and studied both theoretically and experimentally. The two optical methods are compared from two aspects of flow field structure visualization and parameter diagnosis. On this basis, the feasibility of combining the background oriented schlieren with the laser beam profile deformation method to visualize and diagnose the flow field is proposed, and an experimental device is established. In addition, the hydrogen flow field is selected as an empirical example. The experimental results show that the combination of background oriented schlieren and the laser beam profile deformation method may be a reasonable and feasible means for complex flow field structure visualization and parameter diagnosis.

We propose a new scheme for recognizing the topological charge (TC) of orbital angular momentum (OAM) beams using convolutional neural networks (CNN) based on the focusing of cylindrical lenses and the detection of linear photodiode arrays (PDAs). Simulations and experiments are conducted. For the superimposed OAM sets with different TC values and different TC intervals, the effects of atmospheric turbulence disturbances on recognition accuracy are explored separately, where the turbulence disturbances to the superimposed OAM beams are measured by the coherence length r₀. The simulation results show that the recognition accuracy decreases as the turbulence disturbances increase. With 16-unit PDAs, the TC of the superimposed OAM set l∈{±1, ±2, ±3, ±4} can be recognized with 100% accuracy under weak (the coherence length r₀ =16.16 cm ) and intermediate (the coherence length r₀ =10.66 cm ) turbulence disturbances, and above 90% accuracy under a strong (the coherence length 0 r = 4.06 cm ) turbulence disturbance. In the experiment under weak (the coherence length r₀ =13.01 cm) and intermediate (the coherence length r₀ = 8.59 cm) laboratory-simulated turbulence disturbances, with 16-unit PDAs, the recognition accuracy reaches 100% and 99.65%, respectively. The experimental results verify the results of the simulation.

Gait recognition is of great significance in many fields such as human identity recognition, medical rehabilitation, remote monitoring, and so on. In this paper, the gait signal pattern recognition as well as the identity recognition is achieved through machine learning based on the data obtained by phase sensitive time-domain reflectometer ( φ - OTDR). The dynamic phase variation caused by the pedaling directly acting on the fiber was recorded. By extracting the temporal-spatial features of the recorded signal, the recognition accuracy of the two experimenters can be resolved as high as 90% via a convolutional neural network (CNN). It provides an effective solution for human identity identification in perimeter security applications using φ-OTDR.

In recent years, metalens showed their applied potential as depth information sensors. To combine the advantages of active-imaging Lidar and the depth information acquisition ability of metalens, a 3D imaging setup based on metalens array with strong chromatic dispersion was proposed in this paper. When the scene was illuminated by the laser with different wavelengths, the point spread function (PSF) of points in the scene was varied at a certain imaging plane behind the metalens. Only when the point located on the object plane was illuminated by a certain wavelength the optimized PSF could be observed on the imaging plane. Therefore, by measuring the PSFs of a pairs of defocus images induced by the chromatic dispersion of the metalens the absolute distance from metalens to a point on the scene could be obtained. Combined with the clear 2D image, 3D imaging can be realized. When the chromatic dispersion of the metalens is stronger, PSF is more sensitive to the distance change, higher distance measurement resolution can be obtained. The simulation results of metalens with different dispersion shown in the paper proved our assumption.

Nuclear magnetic resonance gyroscopes (NMRGs) have broad application perspectives with the advantages of low cost, low power consumption, miniaturization-ability and high precision. The transverse relaxation rate of noble gas nuclear spins is used to evaluate the performance of vapor cell, which also affects the angle random walk (ARW) of NMRG systems. The inhomogeneity of electronic spin polarization spatial distribution is one of the essential sources of the transverse relaxation rate. In this paper, we study the influence of the pump power and beam diameter in the transverse relaxation rate of noble gas nuclear spins through numerical simulations of electronic spin polarization and experimental measurements of transverse relaxation time. Simulations of the electronic spin polarization spatial distribution are proposed based on the Bloch–Torrey equations. The transverse relaxation time of noble gas nuclear spins under different pump power and beam diameters is measured by the free induction decay (FID) method. Experimental results show that the transverse relaxation rate of nuclear spins increases with pump power. The relaxation rate with a 2.3mm pump beam diameter is larger than with a 1.3mm diameter. Furthermore, we innovatively find that the transverse relaxation rate shows a linear relationship with the electronic spin polarization obtained from the numerical simulation. This work provides a reference for the study of nuclear spin relaxation and the optimization of the parameters of the pump beam in NMRGs.

A wavelength scanning scheme enabled by phase-shifted fiber Bragg grating (PS-FBG) is proposed for gas absorption spectroscopy based on optical frequency comb (OFC). The PS-FBG works as an ultra-narrow bandpass optical filter to generate the gas sensing laser probe. To obtain the gas absorption signal, the electronical frequency beating and lock-in amplification are successively performed with the continuous tuning of the PS-FBG by using a piezo transducer (PZT). The intermediate beat note is monitored in real time for lock-in frequency compensation against the repetition frequency drift of the OFC. A carbon monoxide (CO) sensing system in direct absorption spectroscopy (DAS) configuration is developed based on a free-running fiber laser frequency comb. A triangular-wave PZT driving signal of 5 Hz is used for periodical spectrum scanning. At an intermediate beat note of ~ 50 kHz, the DAS signal is obtained with a lock-in constant of 200 μs. The CO detection limit of 0.356% for an integration time of 0.4 s and the minimum detectable absorbance of ~ 0.0021 are achieved, which indicate a better sensitivity performance.