Images Of Beads Research Articles

A deep learning-based model for defect reorganisation in welding/wire arc additive manufacturing

Welding and Wire Arc Additive Manufacturing (WAAM) are crucial processes in various industries, but defects can significantly impact the quality and integrity of the final product. The study proposes a deep learning-based model for automated defect recognition in WAAM. The primary objective is to develop a robust and accurate system that can identify welding defects using Convolutional Neural Networks (CNN), a deep learning model. The research methodology involves inputting weld bead images into the model, which is then trained to recognise different defects. During the training phase, images depicting various defects were used to train the DL model, enabling it to learn the distinctive features and patterns associated with each type of defect. The trained model was later tested, and it demonstrated an accuracy rate of 88% in classifying welding defects, showcasing its potential for real-time monitoring and quality control in welding/WAAM applications. The findings of this study contribute to the advancement of intelligent manufacturing systems and highlight the effectiveness of deep learning techniques in enhancing the reliability and efficiency of welding processes.

A Study on the Nostalgia aspect of Kwon Oh-soon’ work – Focusing on the image of beads -

A Study on the Nostalgia aspect of Kwon Oh-soon’ work – Focusing on the image of beads -

OpenSTED: open-source dynamic intensity minimum system for stimulated emission depletion microscopy.

Stimulated emission depletion (STED) is a powerful super-resolution microscopy technique that can be used for imaging live cells. However, the high STED laser powers can cause significant photobleaching and sample damage in sensitive biological samples. The dynamic intensity minimum (DyMIN) technique turns on the STED laser only in regions of the sample where there is fluorescence signal, thus saving significant sample photobleaching. The reduction in photobleaching allows higher resolution images to be obtained and longer time-lapse imaging of live samples. A stand-alone module to perform DyMIN is not available commercially. In this work, we developed an open-source design to implement three-step DyMIN on a STED microscope and demonstrated reduced photobleaching for timelapse imaging of beads, cells, and tissue. The DyMIN system uses a fast multiplexer circuit and inexpensive field-programmable gate array controlled by Labview software that operates as a stand-alone module for a STED microscope. All software and circuit diagrams are freely available. We compared time-lapse images of bead samples using our custom DyMIN system to conventional STED and recorded a higher signal when using DyMIN after a 50-image sequence. We further demonstrated the DyMIN system for time-lapse STED imaging of live cells and brain tissue slices. Our open-source DyMIN system is an inexpensive add-on to a conventional STED microscope that can reduce photobleaching. The system can significantly improve signal to noise for dynamic time-lapse STED imaging of live samples.

Smartphone-based optical sectioning (SOS) microscopy with a telecentric design for fluorescence imaging.

We propose a smartphone-based optical sectioning (SOS) microscope based on the HiLo technique, with a single smartphone replacing a high-cost illumination source and a camera sensor. We built our SOS with off-the-shelf optical, mechanical cage systems with 3D-printed adapters to seamlessly integrate the smartphone with the SOS main body. The liquid light guide can be integrated with the adapter, guiding the smartphone's LED light to the digital mirror device (DMD) with neglectable loss. We used an electrically tuneable lens (ETL) instead of a mechanical translation stage to realise low-cost axial scanning. The ETL was conjugated to the objective lens's back pupil plane (BPP) to construct a telecentric design by a 4f configuration to maintain the lateral magnification for different axial positions. SOS has a 571.5µm telecentric scanning range and an 11.7µm axial resolution. The broadband smartphone LED torch can effectively excite fluorescent polystyrene (PS) beads. We successfully used SOS for high-contrast fluorescent PS beads imaging with different wavelengths and optical sectioning imaging of multilayer fluorescent PS beads. To our knowledge, the proposed SOS is the first smartphone-based HiLo optical sectioning microscopy (£1965), which can save around £7035 compared with a traditional HiLo system (£9000). It is a powerful tool for biomedical research in resource-limited areas.

Mechanical-scan-free multicolor super-resolution imaging with diffractive spot array illumination.

Point-scanning microscopy approaches are transforming super-resolution imaging. Despite achieving parallel high-speed imaging using multifocal techniques, efficient multicolor imaging methods with high-quality illumination are currently lacking. In this paper, we present for the first time Mechanical-scan-free multiColor Super-resolution Microscopy (MCoSM) with spot array illumination, which enables mechanical-scan-free super-resolution imaging with adjustable resolution and a good effective field-of-view based on spatial light modulators. Through 100-2,500 s super-resolution spot illumination with different effective fields of view for imaging, we demonstrate the adjustable capacity of MCoSM. MCoSM extends existing spectral imaging capabilities through a time-sharing process involving different color illumination with phase-shift scanning while retaining the spatial flexibility of super-resolution imaging with diffractive spot array illumination. To demonstrate the prospects of MCoSM, we perform four-color imaging of fluorescent beads at high resolution. MCoSM provides a versatile platform for studying molecular interactions in complex samples at the nanoscale level.

A novel variational approach for multiphoton microscopy image restoration: from PSF estimation to 3D deconvolution

In multi-photon microscopy (MPM), a recent in-vivo fluorescence microscopy system, the task of image restoration can be decomposed into two interlinked inverse problems: firstly, the characterization of the point spread function (PSF) and subsequently, the deconvolution (i.e. deblurring) to remove the PSF effect, and reduce noise. The acquired MPM image quality is critically affected by PSF blurring and intense noise. The PSF in MPM is highly spread in 3D and is not well characterized, presenting high variability with respect to the observed objects. This makes the restoration of MPM images challenging. Common PSF estimation methods in fluorescence microscopy, including MPM, involve capturing images of sub-resolution beads, followed by quantifying the resulting ellipsoidal 3D spot. In this work, we revisit this approach, coping with its inherent limitations in terms of accuracy and practicality. We estimate the PSF from the observation of relatively large beads (approximately 1 μm in diameter). This goes through the formulation and resolution of an original non-convex minimization problem, for which we propose a proximal alternating method along with convergence guarantees. Following the PSF estimation step, we then introduce an innovative strategy to deal with the high level multiplicative noise degrading the acquisitions. We rely on a heteroscedastic noise model for which we estimate the parameters. We then solve a constrained optimization problem to restore the image, accounting for the estimated PSF and noise, while allowing a minimal hyper-parameter tuning. Theoretical guarantees are given for the restoration algorithm. These algorithmic contributions lead to an end-to-end pipeline for 3D image restoration in MPM, that we share as a publicly available Python software. We demonstrate its effectiveness through several experiments on both simulated and real data.

Quantitative assessment and monitoring of microplastics and nanoplastics distributions and lipid metabolism in live zebrafish using hyperspectral stimulated Raman scattering microscopy

Microplastics (MP) and nanoplastics (NP) pollutions pose a rising environmental threat to humans and other living species, given their escalating presence in essential resources that living subjects ingest and/or inhale. Herein, to elucidate the potential health implications of MP/NP, we report for the first time by using label-free hyperspectral stimulated Raman scattering (SRS) imaging technique developed to quantitatively monitor the bioaccumulation and metabolic toxicity of MP/NP within live zebrafish larvae during their early developmental stages. Zebrafish embryos are exposed to environmentally related concentrations (3–60 μg/ml) of polystyrene (PS) beads with two typical sizes (2 μm and 50 nm). Zebrafish are administered isotope-tagged fatty acids through microinjection and dietary intake for in vivo tracking of lipid metabolism dynamics. In vivo 3D quantitative vibrational imaging of PS beads and intrinsic biomolecules across key zebrafish organs reveals that gut and liver are the primary target organs of MP/NP, while only 50 nm PS beads readily aggregate and adhere to the brain and blood vessels. The 50 nm PS beads are also found to induce more pronounced hepatic inflammatory response compared to 2 μm counterparts, characterized by increased biogenesis of lipid droplets and upregulation of arachidonic acid detected in zebrafish liver. Furthermore, Raman-tagged SRS imaging of fatty acids uncovers that MP/NP exposure significantly reduces yolk lipid utilization and promotes dietary lipid storage in zebrafish, possibly associated with developmental delays and more pronounced food dilution effects in zebrafish larvae exposed to 2 μm PS beads. The hyperspectral SRS imaging in this work shows that MP/NP exposure perturbs the development and lipid metabolism in zebrafish larvae, furthering the understanding of MP/NP ingestions and consequent toxicity in different organs in living species.

Photophysical image analysis: Unsupervised probabilistic thresholding for images from electron-multiplying charge-coupled devices.

We introduce the concept photophysical image analysis (PIA) and an associated pipeline for unsupervised probabilistic image thresholding for images recorded by electron-multiplying charge-coupled device (EMCCD) cameras. We base our approach on a closed-form analytic expression for the characteristic function (Fourier-transform of the probability mass function) for the image counts recorded in an EMCCD camera, which takes into account both stochasticity in the arrival of photons at the imaging camera and subsequent noise induced by the detection system of the camera. The only assumption in our method is that the background photon arrival to the imaging system is described by a stationary Poisson process (we make no assumption about the photon statistics for the signal). We estimate the background photon statistics parameter, λbg, from an image which contains both background and signal pixels by use of a novel truncated fit procedure with an automatically determined image count threshold. Prior to this, the camera noise model parameters are estimated using a calibration step. Utilizing the estimates for the camera parameters and λbg, we then introduce a probabilistic thresholding method, where, for the first time, the fraction of misclassified pixels can be determined a priori for a general image in an unsupervised way. We use synthetic images to validate our a priori estimates and to benchmark against the Otsu method, which is a popular unsupervised non-probabilistic image thresholding method (no a priori estimates for the error rates are provided). For completeness, we lastly present a simple heuristic general-purpose segmentation method based on the thresholding results, which we apply to segmentation of synthetic images and experimental images of fluorescent beads and lung cell nuclei. Our publicly available software opens up for fully automated, unsupervised, probabilistic photophysical image analysis.

Establishing a reference focal plane using convolutional neural networks and beads for brightfield imaging

Repeatability of measurements from image analytics is difficult, due to the heterogeneity and complexity of cell samples, exact microscope stage positioning, and slide thickness. We present a method to define and use a reference focal plane that provides repeatable measurements with very high accuracy, by relying on control beads as reference material and a convolutional neural network focused on the control bead images. Previously we defined a reference effective focal plane (REFP) based on the image gradient of bead edges and three specific bead image features. This paper both generalizes and improves on this previous work. First, we refine the definition of the REFP by fitting a cubic spline to describe the relationship between the distance from a bead’s center and pixel intensity and by sharing information across experiments, exposures, and fields of view. Second, we remove our reliance on image features that behave differently from one instrument to another. Instead, we apply a convolutional regression neural network (ResNet 18) trained on cropped bead images that is generalizable to multiple microscopes. Our ResNet 18 network predicts the location of the REFP with only a single inferenced image acquisition that can be taken across a wide range of focal planes and exposure times. We illustrate the different strategies and hyperparameter optimization of the ResNet 18 to achieve a high prediction accuracy with an uncertainty for every image tested coming within the microscope repeatability measure of 7.5 µm from the desired focal plane. We demonstrate the generalizability of this methodology by applying it to two different optical systems and show that this level of accuracy can be achieved using only 6 beads per image.

Easily scalable multi-color DMD-based structured illumination microscopy.

Structured illumination microscopy (SIM) achieves super-resolution imaging using a series of phase-shifted sinusoidal illumination patterns to down-modulate high spatial-frequency information of samples. Digital micromirror devices (DMDs) have been increasingly used to generate SIM illumination patterns due to their high speed and moderate cost. However, a DMD micromirror array's blazed grating structure causes strong angular dispersion for different wavelengths of light, thus severely hampering its application in multicolor imaging. We developed a multi-color DMD-SIM setup that employs a diffraction grating to compensate the DMD's dispersion and demonstrate super-resolution SIM imaging of both fluorescent beads and live cells samples with four color channels. This simple but effective approach can be readily scaled to more color channels, thereby greatly expanding the application of SIM in the study of complex multi-component structures and dynamics in soft matter systems.

Fully automated point spread function analysis using PyCalibrate.

Reproducibility is severely limited if instrument performance is assumed rather than measured. Within optical microscopy, instrument performance is typically measured using sub-resolution fluorescent beads. However, the process is performed infrequently as it is requires time and suitably trained staff to acquire and then process the bead images. Analysis software still requires the manual entry of imaging parameters. Human error from repeatedly typing these parameters can significantly affect the outcome of the analysis, rendering the results less reproducible. To avoid this issue, PyCalibrate has been developed to fully automate the analysis of bead images. PyCalibrate can be accessed either by executing the Python code locally or via a user-friendly web portal to further improve accessibility when moving between locations and machines. PyCalibrate interfaces with the BioFormats library to make it compatible with a wide range of proprietary image formats. In this study, PyCalibrate analysis performance is directly compared with alternative free-access analysis software PSFj, MetroloJ QC and DayBook 3 and is demonstrated to have equivalent performance but without the need for user supervision.

Defocused imaging-based quantification of plasmon-induced distortion of single emitter emission

Optical properties of single emitters can be significantly improved through the interaction with plasmonic structures, leading to enhanced sensing and imaging capabilities. In turn, single emitters can act as sensitive probes of the local electromagnetic field surrounding plasmonic structures, furnishing fundamental insights into their physics and guiding the design of novel plasmonic devices. However, the interaction of emitters in the proximity to a plasmonic nanostructure causes distortion, which hinders precise estimation of position and polarization state and is one of the reasons why detection and quantification of molecular processes yet remain fundamentally challenging in this era of super-resolution. Here, we investigate axially defocused images of a single fluorescent emitter near metallic nanostructure, which encode emitter positions and can be acquired in the far-field with high sensitivity, while analyzing the images with pattern matching algorithm to explore emitter-localized surface plasmon interaction and retrieve information regarding emitter positions. Significant distortion in defocused images of fluorescent beads and quantum dots near nanostructure was observed and analyzed by pattern matching and finite-difference time-domain methods, which revealed that the distortion arises from the emitter interaction with nanostructure. Pattern matching algorithm was also adopted to estimate the lateral positions of a dipole that models an emitter utilizing the distorted defocused images and achieved improvement by more than 3 times over conventional diffraction-limited localization methods. The improvement by defocused imaging is expected to provide a way of enhancing reliability when using plasmonic nanostructure and diversifying strategies for various imaging and sensing modalities.

A Study on Changes of Bead Shape and Metal Transfer Mode Affected by Molten Pool Behavior for Welding Position in GMAW Process

In gas metal arc welding, the molten pool is affected by gravity and this affects welding phenomena. This study investigated the changes in bead shape and metal transfer mode caused by the molten-pool behavior. We conducted gas metal arc welding with various welding positions. Images from high-speed camera and signals from current and voltage sensors were used to determine the differences with welding positions. Bead shapes were quantified through transverse cross-sectional bead images and image processing. Molten-pool images from the camera were analyzed through the molten-pool length and distance from the wire position. These were used to explain the differences between transverse and longitudinal cross-sectional bead images with welding positions. Changes in metal transfer based on current and voltage signals were investigated. To determine the exact change in metal transfer, the number of short-circuit transfers and spray transfers within the same time was captured by the welding position camera.

Deep learning for anomaly detection in wire-arc additive manufacturing

Wire-arc additive manufacturing (WAAM) is becoming the most important metal additive manufacturing process in many industries. In this paper, one of the common problems of irregularity in the metal deposition in WAAM has been addressed and solved using machine learning (ML). A deep learning-based convolutional neural network (CNN) was used to classify the two classes of deposited beads, i.e. ‘regular bead’ and ‘irregular bead’. A digital camera was installed with a WAAM setup to obtain the images of beads after deposition. A single layer of deposition was conducted on a substrate using aluminium 5356 alloy filler wire using robotic-controlled gas-metal arc welding (GMAW) setup. The performance of the ML model was validated using classification accuracy and processing time. The developed CNN model was checked with three types of proposed datasets. The dataset containing the training and testing ratio of 60:40 achieved an accuracy of 86.53% and 88.08% with 30 and 60 epochs respectively for testing. The proposed ML model was successful in anomaly detection in the deposited bead of WAAM and hence it helps in improving the quality of deposited layers and mechanical properties of fabricated parts.

Coherent Stokes Raman scattering microscopy (CSRS)

We report the first implementation of laser scanning coherent Stokes Raman scattering (CSRS) microscopy. To overcome the major challenge in CSRS imaging, we show how to suppress the fluorescence background by narrow bandpass filter and a lock-in based demodulation. Near background free CSRS imaging of polymer beads, human skin, onion cells, avocado flesh and the wing disc of a drosphila larva are presented. Finally, we explain and demonstrate numerically that CSRS solves a major obstacle of other coherent Raman techniques by sending a significant part (up to 100%) of the CSRS photons into the backward direction under tight focusing conditions. We believe that this discovery will pave the way for numerous technological advances, e.g., in epi-detected coherent Raman multi-focus imaging, real-time laser scanning based spectroscopy or efficient endoscopy.

<i>In Situ</i> Evaluation of Crack Generation Using Acoustic Emission Technique during Laser Cladding of WC-Co Cemented Carbide via Multi-Beam Laser Metal Deposition

In-situ evaluation of crack generation using acoustic emission (AE) technique as a non-destructive testing method was applied during laser cladding of WC-Co cemented carbide via a multi-beam laser metal deposition (LMD) system. Especially, this study aims to investigate relationships between the laser output and the number of cracks generated in beads of WC-12mass%Co cemented carbide cladded by the LMD system. The number of cracks was evaluated using an AE sensor attached to a substrate. The number of cracks was also evaluated from SEM images of beads. By comparing these results, problems in the both evaluation methods for crack generation in laser cladding were discussed.

Label-free structural and functional volumetric imaging by dual-modality optical-Raman projection tomography

Mesoscale volumetric imaging is of great importance for the study of bio-organisms. Among others, optical projection tomography provides unprecedented structural details of specimens, but it requires fluorescence label for chemical targeting. Raman spectroscopic imaging is able to identify chemical components in a label-free manner but lacks microstructure. Here, we present a dual-modality optical-Raman projection tomography (ORPT) technology, which enables label-free three-dimensional imaging of microstructures and components of millimeter-sized samples with a micron-level spatial resolution on the same device. We validate the feasibility of our ORPT system using images of polystyrene beads in a volume, followed by detecting biomolecules of zebrafish and Arabidopsis, demonstrating that fused three-dimensional images of the microstructure and molecular components of bio-samples could be achieved. Last, we observe the fat body of Drosophila melanogaster at different developmental stages. Our proposed technology enables bimodal label-free volumetric imaging of the structure and function of biomolecules in a large sample.

Overcoming the field-of-view to diameter trade-off in microendoscopy via computational optrode-array microscopy.

High-resolution microscopy of deep tissue with large field-of-view (FOV) is critical for elucidating organization of cellular structures in plant biology. Microscopy with an implanted probe offers an effective solution. However, there exists a fundamental trade-off between the FOV and probe diameter arising from aberrations inherent in conventional imaging optics (typically, FOV < 30% of diameter). Here, we demonstrate the use of microfabricated non-imaging probes (optrodes) that when combined with a trained machine-learning algorithm is able to achieve FOV of 1x to 5x the probe diameter. Further increase in FOV is achieved by using multiple optrodes in parallel. With a 1 × 2 optrode array, we demonstrate imaging of fluorescent beads (including 30 FPS video), stained plant stem sections and stained living stems. Our demonstration lays the foundation for fast, high-resolution microscopy with large FOV in deep tissue via microfabricated non-imaging probes and advanced machine learning.

Precise, wide field, and low-cost imaging and analysis of core-shell beads for digital polymerase chain reaction.

Digital droplet reactors have become a valuable tool for the analysis of single cells, organisms, or molecules by discretising reagents into picolitre or nanolitre volumes. However, DNA-based assays typically require processing of samples on the scale of tens of microlitres, with the detection of as few as one or as many as a hundred thousand fragments. Through the present work, we introduce a flow-focusing microfluidic device that produces 120 picolitre core-shell beads, which are assembled into a monolayer in a Petri dish for visualization and analysis. The bead assembly is subjected to polymerase chain reaction (PCR) amplification and fluorescence detection to digitally quantify the DNA concentration of the sample. We use a low-cost 21-megapixel digital camera and macro lens to capture wide-field fluorescence images with a 10-30 mm2 field-of-view at magnifications ranging from 5× to 2.5×. A customised Python script analysed the acquired images. Our study demonstrates the ability to perform digital PCR analysis of the entire bead assembly through end-point imaging and compare the results with those obtained through RT-qPCR.

Characteristics of the Z-resolution Measurement in the Digital Breast Tomosynthesis

Digital breast tomosynthesis (DBT) offers a choice of image acquisition conditions. The object of observation is the entire detector in the full depth (Z-axis) direction. In this study, we investigated the effect of each acquisition condition on Z-resolution with a micro-metallic sphere (bead) as an imaged object in DBT images. DBT images were acquired in two imaging modes with different angular ranges, varying with the height of the bead, the bead placement position, and the bead diameter. The depth profile of the bead image was obtained, and the Z-resolution was defined by its full width at half maximum. The Z-resolution was compared with each acquisition condition. The Z-resolution values decreased with increasing height from the breast support and decreasing bead diameter. The Z-resolution values were larger at the edges than at the detector center. This study shows that the bead diameter and the geometry of the bead and detector affect the Z-resolution measurement.