Defocus Distance Research Articles

Keratorefractive Lenticule Extraction: Early Postoperative Day 1 Outcomes in 1350 Consecutive Procedures.

To evaluate postoperative visual outcomes on postoperative day 1 (POD1) after lenticule extraction in myopia astigmatism in a large cohort of over 1350 consecutive treatments. A retrospective review chart identified 1357 eyes consecutively treated with SCHWIND ATOS using lenticule extraction for myopic astigmatism who had at least 6 months of follow-up completed and for which POD1 was completely recorded. The standard visual and refractive outcomes, and astigmatism outcomes, were analyzed on POD1. As a sanity check, the stability of defocus and uncorrected distance visual acuity (UDVA) were evaluated from day 1 to 1-year postoperative follow-ups. On POD1, 82% of the eyes achieved an UDVA of 20/20 or better, with 82% of the eyes with postoperative UDVA being the same or better than the preoperative corrected distance visual acuity baseline, and only 2% of eyes lost 2 lines of corrected distance visual acuity. Postoperative spherical equivalent (+0.15 ± 0.35D) was within 0.75 D in 98% of the cases, with excellent stability from 1-week to 1-year postoperative follow-ups (11% eyes changing by 0.5D of defocus). Postoperative refractive astigmatism at POD1 (0.01 ± 0.08 D) was highly accurate, with over 99% of the cases within 0.5 D, and 99% within 5 degrees of the attempted axis. The astigmatic correction index (1.00 ± 0.08) showed 98% of the cases within 10% deviation. In this large cohort of over 1350 cases, lenticule extraction using SCHWIND ATOS is safe and effective on POD1, shows stability from 1-week to 1-year postoperative follow-ups, and suggests accurate and precise corrections for both defocus and astigmatism.

Rapid Autofocus Method Based on LED Oblique Illumination for Metaphase Chromosome Microscopy Imaging System

In clinical practice, microscopes are commonly used for imaging chromosomes to diagnose genetic diseases. Achieving precise and rapid autofocusing is a significant challenge in the advancement of high-throughput chromosome imaging systems. Here, we introduce a rapid autofocus method based on LED oblique illumination for dual-objective configuration in metaphase chromosome imaging system. Our method utilizes a programmable LED array for sample illumination, employing a sequential activation of two LEDs from opposing angles to create oblique illumination. The defocus distance is ascertained through image cross-correlation calculation. Illumination with multiple LEDs in the array is used to acquire bright-field images after completing the focusing. Our method can perform continuous autofocus under a 10× objective and a 100× oil immersion objective, with average focusing errors of 1.29 μm and 0.12 μm, respectively. The total imaging time for a single sample has been significantly reduced from approximately 10 min with conventional methods to just 2 min. This study provides preliminary evidence supporting the viability of developing a next-generation high-throughput chromosome scanner employing a LED array.

Joint formation mechanism and mechanical properties of laser brazed Zn coated steel under different defocusing conditions

Laser brazing, producing a class-A joint surface in automotive, relies on the defocus control to manage laser heating mode and braze performance. This work demonstrates that the laser interaction mechanism transitioned from keyhole welding to conduction brazing as defocus distances increased from −20 to ≥18 mm while keeping other parameters consistent. Uniform brazed joints were achieved at defocuses of +22, +25 and + 30 mm. Increased defocus distance resulted in a wider bead, with a larger laser irradiation area and expanded heat affected zone (HAZ), leading to increased steel melting and more Fe-rich precipitates within the Cu braze. The interfacial reaction layers remained Fe(Si) with increasing thickness as defocus changed from +22 to +30 mm, and two distinct Fe(Si) phases were initially identified. The steel Zn coating evaporated upon direct laser irradiation at upper two regions while participated in interfacial reactions at the weld root. At the weld root, a dramatic phase transition from Zn–Cu to Cu was observed, with liquid Zn(Cu) phases particularly forming in joints with a +30 mm defocus, led to solidification cracks that acted as failure initiation sites during tensile testing. Cracks propagated along the interfacial reaction layer/bead interface, or along large Fe-rich precipitates within the bead. A +30 mm defocus produced a lower hardness HAZ than with a +22 mm defocus, due to the higher content of bainite and tempered martensite resulting from a slower cooling rate. This work provides insights into optimizing laser brazing parameters for Zn-coated steel.

Study on Laser Transmission Welding Technology of TC4 Titanium Alloy and High-Borosilicate Glass.

As the demand for high-performance dissimilar material joining continues to increase in fields such as aerospace, biomedical engineering, and electronics, the welding technology of dissimilar materials has become a focus of research. However, due to the differences in material properties, particularly in the welding between metals and non-metals, numerous challenges arise. The formation and quality of the weld seam are strongly influenced by laser process parameters. In this study, successful welding of high-borosilicate glass to a TC4 titanium alloy, which was treated with high-temperature oxidation, was achieved using a millisecond pulsed laser. A series of process parameter comparison experiments were designed, and the laser welding behavior of the titanium alloy and glass under different process parameters was investigated using scanning electron microscopy (SEM) and a universal testing machine as the primary analysis and testing equipment. The results revealed that changes in process parameters significantly affect the energy input and accumulation during the welding process. The maximum joint strength of 60.67 N was obtained at a laser power of 180 W, a welding speed of 3 mm/s, a defocus distance of 0 mm, and a frequency of 10 Hz. Under the action of the laser, the two materials mixed and penetrated into the molten pool, thus achieving a connection. A phase, Ti5Si3, was detected at the fracture site, indicating that both mechanical bonding and chemical bonding reactions occurred between the high-borosilicate glass and the TC4 titanium alloy during the laser welding process.

Image restoration model for microscopic defocused images based on blurring kernel guidance

Defocus blurring imaging seriously affects the observation accuracy and application range of optical microscopes, and the blurring kernel function is a key parameter for high-resolution image restoration. However, its solving process is complicated and high in computational cost. Image restoration based on most neural networks has high requirements on data sets and the image resolution after restoration is limited because of the lack of quantitative estimation of blurring kernels. In this study, an image restoration method guided by blurring kernel estimation for microscopic defocused images is proposed. First, to reduce the blurring kernel estimation error caused by the positive and negative difference in microscopic defocused imaging, a defocused image classification network is designed to classify the input defocused images with different defocus distances and directions, and its output images are input into the blurring kernel extraction network composed of the feature extraction, correlation, and blurring kernel reconstruction layers. Second, a non-blind defocused image restoration model to restore the high-resolution images is proposed by introducing the blurring kernel extraction module into the restoration network based on U-Net, and the blurring kernel estimation and image restoration losses are jointly trained to realize image restoration guided by blurring kernel estimation. Finally, the experimental results of our proposed method demonstrate significant improvements in both the peak signal-to-noise ratio and structural similarity index measure when compared to other methods.

Multi-field Split Curvature Wavefront Sensing and its Application in the Large Field Survey Telescope

The image quality of a large field survey telescope with a fast focal ratio of the primary mirror is high sensitivity to the optical elements’ misalignments and the primary mirror’s deformations. To maintain good optical performance, the perturbations need to be aligned, and the tomographic reconstruction of these perturbations can be derived from wavefront sensing with multi-field points for alignment. This work undertakes a comprehensive examination related to the implementation and optimization of a multi-field split curvature wavefront sensing system, including defocus distance, signal pre-processing, deblending of overlapped doughnuts, field-dependent correction, and distorted coordinate correction. We also conduct experiments to demonstrate the application and performance of a multi-field split curvature wavefront sensing system in Mephisto. In the context of the decentering of the secondary mirror, the coefficient of determination (R) attains a high value of 0.993, indicating a robust linearity between the coma coefficients and the decentering of the secondary mirror. A comparative analysis of the simulated and experimental results shows that the difference between them is less than 0.1 λ.

Influence of discrete laser surface melting on hot cracking behavior and mechanical properties of Q235 steel

Modern industries demand higher mechanical properties from components. Traditional heat treatment methods often suffer from drawbacks such as high production costs, time-intensive processes, and labor-intensive procedures. The integration of bionic engineering with laser surface melting has given rise to Discrete Laser Surface Melting (DLSM) technology, effectively mitigating these shortcomings. Previous research has substantiated the capability of this technology to significantly enhance various metal properties. This study employed an Nd:YAG pulsed laser for the DLSM treatment of Q235 steel. The discrete laser surface melted (DLSMed) units were prepared on Q235 steel samples. Subsequent analysis utilized scanning electron microscopy (SEM), X-ray diffraction (XRD), and energy-dispersive spectroscopy (EDS) to examine the microstructure and chemical composition of the DLSMed samples. Surface profilometry, hardness testing, and X-ray stress testing were conducted to measure surface roughness, microhardness, and residual stress, respectively. Mechanical properties were evaluated using tensile and impact testing machines. Experimental findings revealed that the DLSMed sample's surface could be divided into three regions: the melting zone, the heat-affected zone (HAZ), and the substrate. The microstructure of the melting zone consisted of refined martensite. The depth and width of the DLSMed units exhibited a certain regularity with variations in the defocus distance. Increasing the defocus distance positively impacted grain refinement and hardness but also led to an increase in surface roughness. Residual stress increased within a specific defocusing range, but beyond this range, its variation lost regularity. Hot cracking behavior within the DLSMed unit encompassed four stages: crack nucleation, propagation, healing, and decomposition. Different defocusing distances and the distribution of DLSMed units significantly affected the tensile strength and impact toughness of the samples. As the defocusing distance increased, the tensile and yield strength of DLSMed samples with transverse units gradually decreased. In contrast, samples with longitudinal units showed a trend of first decreasing and then increasing in tensile and yield strength. The distribution of DLSMed units played a significant role in improving the tensile properties of Q235 steel. The contribution of three factors to the impact performance of DLSMed samples was ranked from high to low as hardness, grain size, and yield strength ratio. The soft phase (substrate) could absorb impact energy and convert it into strain energy stored in the DLSMed sample. Only when the distribution of DLSMed units (hard phases) aligned with the tensile direction could they effectively limit the plastic deformation and enhance the mechanical properties of Q235 steel. However, an increase in defocusing distance led to increased brittleness, deteriorating the mechanical properties of the DLSMed samples. Finally, this study unveiled the strengthening mechanisms of DLSMed samples, providing valuable insights for future applications in enhancing the mechanical properties of metals.

Enhancing Accuracy In 3D Defocusing Particle Tracking Using 3D-Printed Ramp

Three-dimensional defocusing particle tracking velocimetry (3D DPTV) is an emerging tool for investigating fluid motion within microfluidic devices, where observation is typically limited to a single direction. This technique enables the reconstruction of 3D flow fields by analyzing the defocused particle images. The accuracy of 3D DPTV significantly relies on the collection of a proper calibration image stack containing particle images with precisely known defocus distances. One major challenge is to insert a microscale calibration target inside the microfluidic devices due to the small length scales of microfluidic devices. While the alternative approaches such as using tilted glass slides or conducting z-scans of particles fixed at the bottom wall, face several limitations. Here, we present the application of a two-photon polymerization (2PP) based method to manufacture a microscale reference ramp as the calibration target. We demonstrate the fabrication of a precise reference ramp structure that spans the desired calibration range to enhance accuracy in 3D defocusing particle tracking. We experimentally and computationally compare the result with the commonly employed methods to obtain the calibration image stack. The performance of the 2PP-based calibration ramps is compared with the z-scanning method using the General Particle tracking (GPTV). The results demonstrate that the z-scanning method underestimates the defocused distances compared to the ground truth provided by the ramp approach, with the error increasing as the defocused distance grows. The integration of microscale calibration targets using 2PP fabrication method offers a significant advancement in the reliability and accuracy of 3D DPTV measurements in microfluidic systems. By enabling precise, device-specific calibration, this approach can enhance the fidelity of flow field reconstructions, benefiting a wide range of microfluidic applications that rely on detailed velocity field information.

Neural-field-assisted transport-of-intensity phase microscopy: partially coherent quantitative phase imaging under unknown defocus distance

The transport-of-intensity equation (TIE) enables quantitative phase imaging (QPI) under partially coherent illumination by measuring the through-focus intensities combined with a linearized inverse reconstruction algorithm. However, overcoming its sensitivity to imaging settings remains a challenging problem because of the difficulty in tuning the optical parameters of the imaging system accurately and because of the instability to long-time measurements. To address these limitations, we propose and experimentally validate a solution called neural-field-assisted transport-of-intensity phase microscopy (NFTPM) by introducing a tunable defocus parameter into neural field. Without weak object approximation, NFTPM incorporates the physical prior of partially coherent image formation to constrain the neural field and learns the continuous representation of phase object without the need for training. Simulation and experimental results of HeLa cells demonstrate that NFTPM can achieve accurate, partially coherent QPI under unknown defocus distances, providing new possibilities for extending applications in live cell biology.

Depth-of-field extended Fourier ptychographic microscopy without defocus distance priori.

Fourier ptychographic microscopy (FPM) provides a solution of high-throughput phase imaging. Thanks to its coherent imaging model, FPM has the capacity of depth-of-field (DOF) extension by simultaneously recovering the sample's transmittance function and pupil aberration, which contains a defocus term. However, existing phase retrieval algorithms (PRs) often struggle in the presence of a significant defocus. In this Letter, different PRs with embedded pupil recovery are compared, and the one based on the alternating direction multiplier method (ADMM-FPM) demonstrates promising potential for reconstructing highly defocused FPM images. Besides, we present a plug-and-play framework that integrates ADMM-FPM and total variation or Hessian denoiser for pupil function enhancement. Both simulations and experiments demonstrate that this framework enables robust reconstruction of defocused FPM images without any prior knowledge of defocus distance or sample characteristics. In experiments involving USAF 1951 targets and pathologic slides, ADMM-FPM combined with the Hessian denoiser successfully corrected the defocus up to approximately 200 µm, i.e., extending the DOF to 400 µm.

FPM-WSI: Fourier ptychographic whole slide imaging via feature-domain backdiffraction

Fourier ptychographic microscopy (FPM) theoretically provides a solution to the trade-off between spatial resolution and field of view (FOV), and has promising prospects in digital pathology. However, block reconstruction and then stitching has become an unavoidable procedure for reconstruction of large FOV due to vignetting artifacts. This introduces digital stitching artifacts, as the existing image-domain optimization algorithms are highly sensitive to systematic errors. Such obstacles significantly impede the advancement and practical implementation of FPM, explaining why, despite a decade of development, FPM has not gained widespread recognition in the field of biomedicine. We report a feature-domain FPM (FD-FPM) based on the structure-aware forward model to realize stitching-free, full-FOV reconstruction. The loss function is uniquely formulated in the feature domain of images, which bypasses the troublesome vignetting effect and algorithmic vulnerability via feature-domain backdiffraction. Through massive simulations and experiments, we show that FD-FPM effectively eliminates vignetting artifacts for full-FOV reconstruction, and still achieves impressive reconstructions despite the presence of various systematic errors. We also found it has great potential in recovering the data with a lower spectrum overlapping rate, and in realizing digital refocusing without a prior defocus distance. With FD-FPM, we achieved full-color and high-throughput imaging (4.7 mm diameter FOV, 336 nm resolution in the blue channel) free of blocking-and-stitching procedures on a self-developed Fourier ptychographic microscopy whole slide imaging platform. The reported FD-FPM shows the value of FPM for various experimental circumstances, and offers physical insights useful for the developments of models for other computational imaging techniques. The reported platform demonstrates high-quality, high-speed imaging and low cost, and could find applications in many fields of biomedical research, as well as in clinical applications.

Spectral speckle displacement in defocused and tilted imaging systems

Speckle patterns offer valuable insights into the surface characteristics or the characteristics of the light generating the speckle. One possible way to extract this information is via spectral speckle correlation (SSC). The cross-correlation between two speckle fields, generated at different wavelengths, can be used for example to determine the roughness of the illuminated surface. Taking defocused measurements of the surface or measuring on a tilted surface leads to a displacement between the speckle, which in turn affects the cross-correlation and leads to errors in the calculated roughness. In this work we present a model to determine the lateral speckle displacement for a change in wavelength in the case of subjective speckle and defocused, tilted objects. This model is therefore applicable to a wide range of applications and allows to estimate and correct for this speckle displacement. Experimental results show sub-pixel accuracy for object tilts up to ±7° and defocus distances up to ±25 mm.

Microstructure, Variant Selection, and Mechanical Properties of Laser-Welded Ti-4Al-2V Joints

Laser welding of the near α-phase titanium alloy Ti-4Al-2V, used for complex components in the nuclear industry, has been rarely reported. In this study, butt weld joints made of Ti-4Al-2V alloy plates under different parameters, including the laser power, the welding speed, and the defocus distance, were manufactured and analyzed. The results showed that adjusting the combination of 4.2 kW of laser power, a 20 mm/s welding speed, and a −2 mm defocus distance could achieve a penetration depth exceeding 6 mm. Porosity defects were prone to forming in the middle and bottom parts of the fusion zone, due to rapid cooling. The microstructure of the fusion zone was mainly needle-like α martensite, which precipitated in the form of specific clusters. The interior of a cluster was composed of three types of variants with <11−20>/60° phase interfaces to achieve the lower boundary’s energy. Affected by the microstructure and welding defects, the strength of the weld joint was basically similar under different welding conditions, namely about 720 MPa, slightly higher than that of the base metal, while the rupture elongation at breaking decreased by more than 50%. The micro-Vickers hardness of the weld joints was about 50–60 HV higher than that of the base metal, while the impact toughness was about 40 KJ, almost half that of the base metal. This research lays a solid foundation for the engineering application of laser welding of Ti-4Al-2V alloys.

Process optimization and performance verification of CFRP laser surface modification

AbstractAdhesive bonding has gradually replaced mechanical connections as the primary method for composite material bonding, with surface quality being the key factor determining the effectiveness of composite material adhesion. Laser surface treatment, known for its green and precise characteristics, has garnered widespread attention from both academia and industry. However, the process parameters of laser surface treatment are complex, efficiency is low, and the mechanism remains unclear. This paper, based on the response surface methodology (RSM), investigates the influence of laser secondary parameters on the resin removal rate of carbon fiber reinforced polymer/plastic (CFRP) surfaces and the impact of defocusing on surface treatment efficiency. The results indicate that when the total laser energy density (Et) is 69 J/cm2, the overlap ratio in the warp direction (OLs) is 76%, and the overlap ratio in the weft direction (OLh) is 85%, the surface tensile shear strength increases by 90.59% and 27.77% compared to the original surface and polished surface, respectively. When the defocus distance is +8 mm, the efficiency of laser surface treatment is 2.25 times that of non‐defocus, with minimal impact on mechanical performance. The surface treatment mechanism, with changes in Et, primarily involves the pure thermal decomposition of the resin matrix in an oxygen‐free environment, oxidative decomposition, and the synergistic action of the decomposition products' oxidation. Through the coupled optimization of laser process parameters, this work enhances the efficiency and effectiveness of composite material adhesive surface treatment, promoting the engineering application of laser technology.Highlights By establishing a response surface model, the study elucidated the impact patterns of laser secondary parameters (overlap rates in the warp and weft directions, total energy density) on the resin removal rate of CFRP surfaces. Based on the laser surface treatment phenomena at different total energy densities, the study analyzed and revealed the mechanism of near‐infrared (1064 nm) laser resin removal from CFRP surfaces. By increasing the spot size through defocusing, the efficiency of laser surface treatment was enhanced, and the impact on the interfacial properties of CFRP after defocused treatment was validated.

Raman signal optimization based on residual network adaptive focusing

Due to its high sensitivity and specificity, Micro-Raman spectroscopy has emerged as a vital technique for molecular recognition and identification. As a weakly scattered signal, ensuring the accurate focus of the sample is essential for acquiring high quality Raman spectral signal and its analysis, especially in some complex microenvironments such as intracellular settings. Traditional autofocus methods are often time consuming or necessitate additional hardware, limiting real-time sample observation and device compatibility. Here, we propose an adaptive focusing method based on residual network to realize rapid and accurate focusing on Micro-Raman measurements. Using only a bright field image of the sample acquired on any image plane, we can predict the defocus distance with a residual network trained by Resnet50, in which the focus position is determined by combining the gradient and discrete cosine transform. Further, detailed regional division of the bright field map used for characterizing the height variation of actual sample surface is performed. As a result, a focus prediction map with 1μm accuracy is obtained from a bright field image in 120 ms. Based on this method, we successfully realize Raman signal optimization and the necessary correction of spectral information. This adaptive focusing method based on residual network is beneficial to further enhance the sensitivity and accuracy of Micro-Raman spectroscopy technology, which is of great significance in promoting the wide application of Raman spectroscopy.

Machine vision-based automatic focusing method for robot laser welding system

Background Defocus distance is a critical parameter in laser welding, especially when encountering changes in the contour of the welding surface. This paper proposed an automated focusing method to address the challenging issue of accurately adjusting the defocus distance. Methods The proposed method involves several steps. Firstly, a clarity evaluation function based on the Kirsch operator is employed to calculate real-time image clarity of the welding surface captured by the machine vision system. Next, an improved Canny edge detection algorithm is applied to identify the edge contours of the welding surface, from which their central points are extracted. Finally, automatic focusing is achieved by employing a variable step-size hill-climbing algorithm to search for the focal plane. Results To verify the applicability of the automatic focusing method proposed for welding the solder ring, a robot laser welding system was designed and constructed. Experimental results show that the positioning error of the robot after automatic focusing is within ±0.4 mm. The average time required for a single automatic focusing process is 16.27 s. These results demonstrated the successful accomplishment of automatic adjustment and control of the focal length. Conclusions The machine vision-based automatic focusing method proposed in this paper enhances the consistency of the robot’s position after automatic focusing in robot laser welding systems. It elevates the level of automation in the welding process and provides an efficient solution for accurately adjusting the welding focal distance during the laser welding process.

Effect of laser defocusing on duplex stainless steels for higher scan speeds in Laser-Based Powder Bed Fusion (PBF-LB/M)

Duplex stainless steels (DSS) employ a bi-modal microstructure consisting of equal parts ferrite and austenite. However, when processed via laser-based powder bed fusion (PBF-LB/M) the high cooling rates lead to a dominantly ferritic microstructure, thus making a post-process heat treatment necessary. This sparks the interest in accelerating the manufacturing time by increasing scan speeds. Defocusing of the laser beam offers the possibility to alter the melt pool morphology and thereby the melting mode. Therefore, this study presents the influence of the defocusing distance for PBF-LB/M manufactured DSS parts. The melt pool becomes shallower and wider as the defocus distance increases, but also has a more uniform shape at high scan speeds compared to the focused beam. Furthermore, defocusing of the laser beam results in denser parts at scan speeds of up to 1800 mm/s, potentially facilitating higher build rates. The duplex microstructure could be restored by a post process heat-treatment.

Fast digital refocusing Fourier ptychographic microscopy method based on convolutional neural network.

Fourier ptychographic microscopy (FPM) is used to achieve high resolution and a large field of view. However, traditional FPM image reconstruction methods often yield poor image quality when encountering out-of-focus issues during reconstruction. Therefore, this study proposes a defocus-distance regression network based on convolutional neural networks. In an experimental validation, the root-mean-square error calculated from 1000 sets of predicted and true values was approximately 6.2 µm. The experimental results suggest that the proposed method has good generalization, maintains high accuracy in predicting defocus distances even for different biological samples, and extends the imaging depth-of-field of the FPM system by a factor of more than 3.

Experimental validation and mathematical simulation for laser protection performance of light field imaging

Photoelectric imaging systems typically employ a focal plane detector structure, rendering them vulnerable to laser damage. Laser damage can severely impair or even completely deprive the information acquisition capability of photoelectric imaging systems. A laser damage protection method based on a microlens array light field imaging system is proposed to prevent photoelectric imaging systems from laser damage. The technique utilizes the light field modulation effect of the microlens array to homogenize the spot energy, thereby reducing the maximum single-pixel receiving power at the image sensor. The method’s effectiveness has been verified through numerical simulations and experimental validation. First, the laser transmission theoretical model of light field imaging is proposed. Then an experimental setup is established, and measurements are conducted to capture the spot profiles and intensity distributions on the imaging plane across various defocus distances. Finally, the impact of the propagation distance on the maximum single-pixel receiving power and suppression ratio of the light field imaging system is experimentally measured. The simulation and experimental results indicate that, with the proposed method, the energy suppression ratio can easily reach two orders of magnitude, significantly reducing the probability of laser damage in photoelectric imaging systems.

High-precision microscopic autofocus with a single natural image.

In industrial microscopic detection, learning-based autofocus methods have empowered operators to acquire high-quality images quickly. However, there are two parts of errors in Learning-based methods: the fitting error of the network model and the making error of the prior dataset, which limits the potential for further improvements in focusing accuracy. In this paper, a high-precision autofocus pipeline was introduced, which predicts the defocus distance from a single natural image. A new method for making datasets was proposed, which overcomes the limitations of the sharpness metric itself and improves the overall accuracy of the dataset. Furthermore, a lightweight regression network was built, namely Natural-image Defocus Prediction Model (NDPM), to improve the focusing accuracy. A realistic dataset of sufficient size was made to train all models. The experiment shows NDPM has better focusing performance compared with other models, with a mean focusing error of 0.422µm.