Image Acquisition Research Articles

Recent Advancements in Machine Vision Systems for Industrial Defect Detection: A Review

Manufacturing, technology, and society experienced a number of changes throughout the Industrial Revolution. This paper presents a comprehensive review of advancements in Machine Vision Systems (MVS) for smart manufacturing and industrial quality control inspections, drawing from a range of studies that highlight advancements in optical systems, image acquisition techniques, and the pivotal role of deep learning methodologies. With the advent of AI and deep learning, MVS have achieved unprecedented levels of accuracy, speed, and versatility. It addresses key components of visual inspection systems, including optical illumination, image acquisition, and image processing, examining their impact on detection accuracy, efficiency, and robustness. The review identifies prevalent methods and future trends, offering insights for researchers and practitioners in manufacturing, computer vision, and quality control. Ultimately, this review aims to present the recent trends in the field of machine vision and image processing for defect detection in industrial applications, along with highlighting the research gaps for future work. Manufacturing processes inherently transform raw materials into finished products through a series of intricate operations that significantly impact the production line's overall efficiency and output quality. Defects and inconsistencies can lead to substantial economic losses, reputational damage, and potential safety hazards. The integration of machine vision systems has significantly enhanced defect detection in industrial manufacturing by improving efficiency, quality, and reliability. MVS have wide applications across industries, namely, automotive industry, electronics manufacturing, food and beverage industry, pharmaceutical industry and textile industry. Despite the numerous benefits of MVS, several challenges exist related to technical constraints, data requirements, adaptability and generalisation, and computational resources. The review concluded that Machine Vision Systems (MVS) have emerged as a critical component of modern industrial quality control, offering unparalleled capabilities for real-time monitoring, defect detection, and process automation. The integration of Artificial Intelligence (AI) and deep learning has further enhanced the performance and versatility of MVS, enabling them to tackle complex inspection tasks with remarkable accuracy and efficiency.

Chronic lymphocytic leukemia (CLL) screening and abnormality detection based on multi-layer fluorescence imaging signal enhancement and compensation

PurposeFluorescence in situ hybridization (FISH) plays a critical role in cancer screening but faces challenges in signal clarity and manual intervention. This study aims to enhance FISH signal clarity, improve screening efficiency, and reduce false negatives through an automated image acquisition and signal enhancement framework.MethodsAn automated workflow was developed, integrating a dynamic signal enhancement method that optimizes global and local features. An improved Cycle-GAN network was introduced, incorporating residual connections and layer-wise supervision to accurately model and compensate for complex signal characteristics. Key metrics such as signal brightness, edge gradients, contrast improvement index (CII), and structural similarity index (SSIM) were used to evaluate performance.ResultsThe proposed method increased weak signal brightness by 49.02%, edge gradients by 48.61%, and CII by 32.52%. The SSIM reached 0.996, indicating high fidelity to original signals.ConclusionVisual analysis demonstrated clearer, more continuous, and uniform fluorescence signals, effectively mitigating fragmentation and uneven distribution. These improvements reduced false negatives and enhanced genomic abnormality detection accuracy. The proposed method significantly improves FISH signal clarity and stability, providing reliable support for cancer screening, genomic abnormality detection, molecular typing, prognosis evaluation, and targeted treatment planning.

A novel optical multi-image data protection scheme based on phase-truncated asymmetric ciphers and 2D-CSCCM

Abstract Existing remote sensing platforms often overlook the security of the remote sensing image acquisition process, creating opportunities for potential attackers to exploit. To address this issue, this paper proposes a novel optical encryption framework (ECC_DPM) based on elliptic curves and phase truncation to ensure the secure transmission of optical keys. Furthermore, an optical encryption scheme is designed based on ECC_DPM to enable synchronous secure encryption during the remote sensing data acquisition stage, ensuring that remote sensing data remains in a ciphertext state from acquisition to storage and distribution. Finally, the proposed framework and encryption scheme are evaluated for security using a threat model and simulation experiments. The experimental results demonstrate that the encryption scheme effectively resists various statistical analysis attacks.

Hip structural analysis parameters are not associated with the risk of postmenopausal female second hip fracture: a retrospective study

PurposePostmenopausal female patients with a history of a single hip fracture are at higher risk of a second fracture. The poorer clinical outcomes of this patient group warrants evaluating the risk of experiencing a second hip fracture. Therefore, this study aimed to investigate the effectiveness of hip structural analysis (HSA) in assessing the risk of second hip fracture in postmenopausal females.MethodsThis retrospective analysis included 188 patients selected from the Chinese Second Hip Fracture Evaluation (ClinicalTrials.gov identifier: NCT03461237, first registration/posted date: 09/03/2018). They were divided into the second hip fracture (35 cases, with a mean age of 79.33 ± 7.70 years) and the control group (153 cases, with a mean age of 73.41 ± 9.56 years). Parker Mobility Score were determined via telephone follow-up, and two computed tomography scanners were used for images acquisition. All HSA and areal bone mineral density (aBMD) parameters were calculated through Mindways QCTPRO software.ResultsThe refracture group showed increased age, decreased cross-sectional area, total hip aBMD, trochanteric aBMD, and intertrochanteric aBMD (p < 0.05). Total hip and intertrochanteric aBMD have a protective effect on the occurrence of a second hip fracture in postmenopausal women, with odd ratios of 0.61 and 0.57, respectively (p < 0.05). Incorporating HSA parameters into the baseline model (used age, type 2 diabetes mellitus, and the PMS as parameters, AUC = 0.729) does not significantly improve the performance of second hip fracture prediction (AUC = 0.748, p < 0.05 in Delong’s test).ConclusionBased on our findings, HSA does not statistically correlate with the incidence of second hip fracture in postmenopausal women. Incorporating HSA parameters into the (baseline) model does not significantly improve the pridictive capabilities.

Classification of OCT Images of the Human Eye Using Mobile Devices

The aim of this study was to develop a mobile application for Android devices dedicated to the classification of pathological changes in human eye optical coherence tomography (OCT) B-scans. The classification process is conducted using convolutional neural networks (CNNs). Six models were trained during the study: a simple convolutional neural network with three convolutional layers, VGG16, InceptionV3, Xception, Joint Attention Network + MobileNetV2 and OpticNet-71. All of these models were converted to TensorFlow Lite format to implement them into a mobile application. For this purpose, three models with the best parameters were chosen, taking accuracy, precision, recall, F1-score and confusion matrix into consideration. The Android application designed for the classification of OCT images was developed using the Kotlin programming language within the Android Studio integrated development environment. With the application, classification can be performed on an image chosen from the user’s files or an image acquired using the photo-taking function. The results of the classification are displayed for three neural networks, along with the respective classification times for each neural network and the associated image undergoing the classification task. The mobile application has been tested using various smartphones. The testing phase included an evaluation of image classification times and score accuracy, considering factors such as image acquisition method, i.e., camera or gallery.

Enhancing Image Reconstruction Method in High-Frequency Electric Field Visualization Systems Using a Polarized Light Image Sensor

This paper introduces an image processing method, used to achieve uniform sensitivity across the imaging plane in a high-frequency electric field imaging system, that employs an electro-optical crystal and a polarization image sensor. The polarization pixels have two polarization directions, 0° and 90°, in pairs, and, conventionally, their difference is computed first. In contrast, this study proposes a method to separate each polarization image, perform pixel completion, and subsequently perform intensity correction. The proposed method was demonstrated to improve field distribution images acquired using 36 GHz and 30 GHz input signals for a microstrip line and patch antenna, respectively. From the measurement results of the microstrip line, the application of the proposed method reduced the electric field fluctuations on the line from 3.1 dB to 1.5 dB. This image-processing method can be applied sequentially during image acquisition, making it suitable for the real-time imaging of electric fields.

Non-invasive and fully two-dimensional quantitative visualization of transparent flow fields enabled by photonic spin-decoupled metasurfaces

Transparent flow field visualization techniques play a critical role in engineering and scientific applications. They provide a clear and intuitive means to understand fluid dynamics and its complex phenomena, such as laminar flow, turbulence, and vortices. However, achieving fully two-dimensional quantitative visualization of transparent flow fields under non-invasive conditions remains a significant challenge. Here, we present an approach for achieving flow field visualization by harnessing the synergistic effects of a dielectric metasurface array endowed with photonic spin-decoupled capability. This approach enables the simultaneous acquisition of light-field images containing flow field information in two orthogonal dimensions, which allows for the real-time and quantitative derivation of multiple physical parameters. As a proof-of-concept, we experimentally demonstrate the applicability of the proposed visualization technique to various scenarios, including temperature field mapping, gas leak detection, visualization of various fluid physical phenomena, and 3D morphological reconstruction of transparent phase objects. This technique not only establishes an exceptional platform for advancing research in fluid physics, but also exhibits significant potential for broad applications in industrial design and vision.

Abdominal ectopic pregnancy in a mixed-breed dog

An intact, ~ 4-6 year female, mixed-breed dog was presented to an academic spay-and-neuter clinic for a routine ovariohysterectomy and transitioned to an abdominal exploratory after palpation of 2 intraabdominal, firm masses during surgical preparation. Intraoperatively, the masses appeared to arise from the small intestine with intimate association to the surrounding mesentery. Due to the level of communication between the masses and the small intestine, euthanasia and postmortem examination were elected. At postmortem evaluation, 3 highly vascularized and encapsulated masses were within the small intestinal mesentery; when excised, each contained a well-developed fetus and 2 were in the early stages of mummification. The largest fetus was free within moderate serosanguinous fluid, and had a thick vascular structure wrapped around the left forelimb. Histologically, this structure was composed of 2 vessels, consistent with an umbilical vein and artery. The fetal capsule contained moderate number of multinucleated basophilic cells, suspicious for placental trophoblasts. The uterus and ovaries had no gross lesions. Although limited in utility due to the postmortem and ex-vivo nature of image acquisition, computed tomography displayed appropriate late-pregnancy development of fetal skeletons and faint, mineral attenuating structures within the maxillary and mandibular alveoli consistent with developing dentition. Some changes consistent with fetal death were noted in the smallest fetal skeleton. The diagnosis was a full-term abdominal ectopic pregnancy. This case report describes a canine abdominal ectopic pregnancy diagnosed during postmortem examination, with additional characterization using advanced imaging and histopathology.

Considerations and recommendations from the ISMRM diffusion study group for preclinical diffusion MRI: Part 2-Ex vivo imaging: Added value and acquisition.

The value of preclinical diffusion MRI (dMRI) is substantial. While dMRI enables in vivo non-invasive characterization of tissue, ex vivodMRI is increasingly being used to probe tissue microstructure and brain connectivity. Ex vivo dMRI has several experimental advantages including higher SNR and spatial resolution compared to in vivo studies, and enabling more advanced diffusion contrasts for improved microstructure and connectivity characterization. Another major advantage of ex vivo dMRI is the direct comparison with histological data, as a crucial methodological validation. However, there are a number of considerations that must be made when performing ex vivo experiments. The steps from tissue preparation, image acquisition and processing, and interpretation of results are complex, with many decisions that not only differ dramatically from in vivo imaging of small animals, but ultimately affect what questions can be answered using the data. This work represents "Part 2" of a three-part series of recommendations and considerations for preclinical dMRI. We describe best practices for dMRI of ex vivo tissue, with a focus on the value that ex vivo imaging adds to the field of dMRI and considerations in ex vivo image acquisition. We first give general considerations and foundational knowledge that must be considered when designing experiments. We briefly describe differences in specimens and models and discuss why some may be more or less appropriate for different studies. We then give guidelines for ex vivo protocols, including tissue fixation, sample preparation, and MR scanning. In each section, we attempt to provide guidelines and recommendations, but also highlight areas for which no guidelines exist (and why), and where future work should lie. An overarching goal herein is to enhance the rigor and reproducibility of ex vivo dMRI acquisitions and analyses, and thereby advance biomedical knowledge.

Enhanced Blood Vessel Visualization and Accelerated Image Acquisition Using Spiral MRA in Moyamoya Disease: A Comparative Study with Cartesian MRA.

To compare the blood vessel visualization with spiral MRA (MRAspiral) and compressed SENSE accelerated Cartesian MRA (MRACS) in moyamoya disease (MMD) patients, with digital subtraction angiography (DSA) as the reference standard. We prospectively collected MRAspiral with different acquisition windows (τ = 4, 6, 10 ms), MRACS and DSA in MMD patients. Contrast-to-noise ratio (CNR) was measured in the M1, M2, M3, and M4 segments of the middle cerebral artery (MCA) for each MRA sequence. Vessel visualization of the distal MCA, leptomeningeal artery (LMA) collaterals, distal external carotid artery (ECA), and internal carotid artery (ICA) steno-occlusion was qualitatively analyzed using 3- and 4-point likert scales compared to DSA. A linear fixed-effects model was used to determine differences among the four sequences. A total of 98 hemispheres from 55 MMD patients were included. CNR in the M2, M3 and M4 segments of the MCA was not significantly different between MRACS and MRAτ4 or MRAτ6, but it was significantly higher in MRACS than MRAτ10 (M2: P < 0.001, M3: P < 0.001, M4: P = 0.013). MRAspiral sequences provided better visualization of the distal MCA, LMA collaterals and distal ECA compared to MRACS (all P < 0.001). MRAspiral offers improved vessel visualization in distal arteries with adequate image quality for patients with MMD. Compared to MRACS, MRAspiral can reduce scan time by 32.31% when the τ value is set to 6 ms, while also providing superior image quality. Spiral MRA performs well in visualizing collateral vessels in moyamoya disease with shorter scan time.

Hand–Eye Separation-Based First-Frame Positioning and Follower Tracking Method for Perforating Robotic Arm

Hand–Eye Separation-Based First-Frame Positioning and Follower Tracking Method for Perforating Robotic Arm

High dynamic range fringe projection profilometry with binary fringes differencing demodulation

Fringe projection profilometry (FPP) using gray-scale patterns like sinusoidal stripes often encounters challenges for high dynamic range (HDR) applications. These challenges arise due to the presence of shiny surfaces or abrupt changes in reflectivity, leading to either reconstruction failure or substantial errors. We present a robust modulation and reconstruction method that divides the grayscale sinusoidal fringe into eight binary fringes and employs a differencing approach to mitigate HDR surface interference and enable demodulation, significantly reducing vulnerability to reflectivity fluctuations. The binary fringes differencing demodulation (BFDD) technique is presented to extract the relevant binary patterns from the camera images and integrate them into an 8-bit HDR gray stripe modulation. Binary modulation allows for simultaneous projection and image acquisition in just 0.2 seconds. The experimental results showed that this method achieves a remarkably high coverage rate in high-reflectivity 3D reconstruction. The method can also effectively mitigate negative impact due to abrupt changes in surface reflectivity, resulting in near-perfect ground truth consistency.

Use of Point-of-Care Ultrasound in Home vs Non-home Medical Care Physicians in Japan.

Use of Point-of-Care Ultrasound in Home vs Non-home Medical Care Physicians in Japan.

Image Deblurring Algorithm Incorporating Self-Attention Mechanism

The acquisition of clear images is a critical aspect in various fields including computer vision, aerial detection, and medical imaging. The issue of image blur caused by object motion poses a challenge in obtaining clear images. To address this, an improved AT-DGAN network model is proposed in this paper. This model integrates the pyramid generator module of the DeblurGAN-v2 network with a self-attention mechanism. The feature pyramid is employed for image feature extraction and representation, while the self-attention mechanism dynamically adjusts the weight of important features in each pyramid layer and performs weighted fusion, thereby compensating for the information loss during feature extraction in the feature pyramid network. Additionally, a hinge loss function is designed for the proposed model to balance the discriminator and the generator, enhancing the stability and training efficiency of the generative adversarial network. The experimental results show that compared to other algorithms of the same type, this improved algorithm has increased the Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) of restored images by 0.58[Formula: see text]dB and 1.5%, respectively.

Emergency Dyspnea in 258 Cats: Insights from the French RAPID CAT Study

Dyspnea is common in feline emergency medicine. Dyspneic cats are often unable to tolerate additional diagnostic tests upon admission. The aim of this study was to determine the relative frequency of the causes of severe dyspnea in cats and to identify clinical indicators that could guide immediate management. Records of 312 severely dyspneic cats admitted over a 5-year period to our institution’s emergency and critical care unit were retrospectively reviewed. The cause of dyspnea was identified for 258 cats, representing 83% of the cases. This study focused solely on these 258 cats. Respiratory causes accounted for 33% of cases, followed by cardiac causes (25%) and both traumatic and neoplastic causes (21% each). Pleural effusion was present in 39% of the cats, with various origins: cardiac (38%), respiratory (pyothorax, feline infectious peritonitis, 30%), neoplastic (23%), and traumatic (9%). Male cats were significantly overrepresented in the cardiac and respiratory groups. The median age differed according to the cause (2 years for traumatic causes, 6 years for respiratory causes, and approximately 11 years for cardiac and neoplastic causes). Cats with cardiac disease were significantly more often hypothermic. Hyperthermia was not systematically indicative of an infectious process. No clinical element (temperature, heart rate, cardiac auscultation, etc.) allowed for predicting the cause of dyspnea. In-hospital mortality was 44%, varying according to the cause (78% for neoplastic causes, 40% for traumatic causes, 37% for cardiac causes, and 30% for respiratory causes). In conclusion, the four main categories of causes were represented in approximately equal proportions, which contrasts with the results of the British RAPID CAT study, where cardiac causes accounted for nearly two-thirds of dyspnea cases. In the dyspneic cats in this series, the clinical presentation did not allow clinicians to identify a particular cause. This highlights the importance of non-invasive examinations such as POCUS (point-of-care ultrasound), which has become a routine test in emergency medicine. However, this importance should not lead to underestimating the continued value of radiography when the cat’s clinical condition allows for image acquisition.

Value of clinical review for AI-guided deep vein thrombosis diagnosis with ultrasound imaging by non-expert operators

Deep vein thrombosis (DVT) carries high morbidity, mortality, and costs globally. Point of care ultrasound (POCUS) image acquisition by non-ultrasound-trained providers, supported by an AI-based guidance and remote image review system, is believed to improve the timeliness and cost-effectiveness of diagnosis. We examine a database of 381 patients with suspected DVT who underwent an AI-guided ultrasound scan by a non-ultrasound-trained nurse and an expert sonographer-performed standard compression ultrasound scan. Each AI-guided scan was reviewed remotely by blinded radiologists or blinded independent POCUS-certified American Emergency Medicine (EM) physicians. Remote reviewer and standard scan diagnoses were compared. The primary endpoint is AI-guidance system sensitivity with clinician review; secondary endpoints include specificity, positive predictive value, negative predictive value, image quality, inter-observer image quality, and vein compressibility agreement. Data was analysed through the bootstrapping method, bootstrapping with a second reader for each scan, and a majority voting system. Eighty percent (n = 304) of scans were of sufficient diagnostic quality. Radiologist reviewer sensitivity ranged from 90%–95%, specificity from 74%–84%, NPV from 98%–99%, PPV from 30%–42%, and potential expert-led ultrasound scans avoided from 39%–50%. Inter-observer agreement for image quality was 0.15 and for compressibility 0.61. EM reviewer sensitivity ranged from 95%–98%, specificity from 97%–100%, NPV was 99%, PPV from 81%–100%, and potential expert-led ultrasound scans avoided from 29%–38%. Inter-observer agreement for image quality was 0.59 and for compressibility 0.67. Diagnosing lower extremity DVT through AI-guided image acquisition with clinician review is feasible. Performance is influenced by reviewer expertise. We find potential positive impacts on health economics, including safely avoiding expert-led ultrasound scans.

An automated and high-throughput approach for enhanced precision of adenoviral titering.

An automated and high-throughput approach for enhanced precision of adenoviral titering.

Protocol for the three-dimensional analysis of rodent skeletal muscle.

Protocol for the three-dimensional analysis of rodent skeletal muscle.

Enhance registration precision of transmission breast images utilizing improved Levenberg-Marquardt optimization algorithm with normalized cross-correlation.

Enhance registration precision of transmission breast images utilizing improved Levenberg-Marquardt optimization algorithm with normalized cross-correlation.

Protocol for functional screening of CFTR-targeted genetic therapies in patient-derived organoids using DETECTOR deep-learning-based analysis.

Protocol for functional screening of CFTR-targeted genetic therapies in patient-derived organoids using DETECTOR deep-learning-based analysis.