3D Image Analysis Research Articles

Measuring pore sizes of nonwoven geotextiles by X-ray CT images: a comparative evaluation

This paper focuses on evaluating the pore sizes of four nonwoven geotextile specimens with different pore size distribution techniques. The paper introduces a bubble analysis technique to find the two-dimensional (2D) pore size distribution of geotextile specimens. A new method is also introduced to determine the three-dimensional (3D) constriction sizes of nonwoven geotextiles from X-ray CT images. The digital image concept and image processing methods for the 2D image analysis and 3D constriction size measurement of geotextile specimens are explained. Finally, the paper compares the 3D constriction sizes with those measured from 2D image analysis, mercury intrusion porosimetry, manufacturer-reported opening size values based on the wet sieving method, and pore sizes calculated from theoretical equations. The results show that the largest pore opening sizes measured from the 3D constriction size method are in good agreement with the theoretical equations. The values measured from the wet sieving method were found to be the smallest, indicating the limitations of the wet sieving method to determine the largest pore opening size of nonwoven geotextiles. As it is the constriction size which must control filtration behaviour, the 3D constriction size method is recommended for assessing the reference porometry.

Temporomandibular joint CBCT image segmentation via multi-view ensemble learning network.

Accurate segmentation of the temporomandibular joint (TMJ) from cone beam CT (CBCT) images holds significant clinical value for diagnosing temporomandibular joint osteoarthrosis (TMJOA) and related conditions. Convolutional neural network-based medical image segmentation methods have achieved state-of-the-art performance in various segmentation tasks. However, 3D medical images segmentation requires substantial global context and rich spatial semantic information, demanding much more GPU memory and computational resources. To address these challenges in 3D medical image segmentation, we propose a novel network- the MVEL-Net (Multi-view Ensemble Learning Network) for TMJ CBCT image segmentation. By resampling images along three dimensions, we generate multiple weak learners with different spatial semantic information. A subsequent strong learning network effectively integrates the outputs from these weak learners to achieve more accurate segmentation results. We evaluated our network model using a clinical dataset comprising 88 subjects with TMJ CBCT images. The average Dice similarity coefficient (DSC) was 0.9817 ± 0.0049, the average surface distance was 0.0540 ± 0.0179mm, and the 95% Hausdorff distance was 0.1743 ± 0.0550mm. Our proposed MVEL-Net demonstrates excellent segmentation performance on TMJ from CBCT images, while using fewer GPU memory resources compared to other 3D networks. The effectiveness of this method in capturing spatial context could be leveraged for tasks like organ segmentation from volumetric scans. This may facilitate wider adoption of AI-based solutions for automated analysis of 3D medical images.

In-depth morphological and functional analysis of ATTR-CA myocardial cells and niche by electron microscopic 3D reconstruction

Abstract Background Cell-cell interactions and interplay with extracellular matrix are known to be important in the steady state and pathological conditions, including wild-type transthyretin cardiac amyloidosis (wtATTR-CA). However, alterations of the microstructure of cardiomyocytes (CMs) and cardiac niches have not been well analyzed. Normal histological staining is usually limited in resolution, and 2D electron microscopy cannot capture stereoscopic images or the distribution of target particles and organelles in intra- and extracellular 3D space, so novel methods are being attempted. Methods and Results Using Serial Block Face-Scanning Electron Microscopy (SBF-SEM) and a 3D structural reconstruction system, we for the first time investigated the microstructure of human myocardial tissue from control hypertrophied heart (hypertensive heart disease) and wtATTR-CA patients. The microstructure of myocardial niche, cell morphology/numbers, and the distribution of intracellular organelles such as intercalated discs (ICDs) were compared to control endomyocardial biopsy (EMB) samples (2 controls vs 5 wtATTR-CA patients). In all wtATTR-CA samples, amyloid fibers were deposited originating around the capillary structures with various electron densities, compressing myocardial cells not only from the periphery but also dividing them from the center of the cell body, and CMs showed degeneration and dwarfing according to the severity of amyloid fiber deposition (Figure 1). The capillary structures showed degeneration and narrowing with stratified basement membrane, and rarefaction (3378±216.5 vs 1448±116.8/mm2, p<0.0001). Notably, the stromal cells (pericytes, fibroblasts, macrophages) were more likely buried by amyloid fibers than CMs, and their numbers were also reduced, indicating a loss of adjacent and intercellular interaction with capillaries and CMs (Fig. 2, 1892±194.4 vs 1189±107.4 mm2, p=0.0043). This method can also be used to analyze intracellular organelles and microstructures in CMs. In ATTR-CA, morphological abnormalities were observed in ICDs (Figure 3) and mitochondria. Conclusion In the ATTR-CM patients, there was a significant decrease in the number of capillaries in accordance with the deposition of myocardial amyloid fibers, as well as a decrease in stromal cells. These results support the hypothesis that amyloid fibril deposition not only causes chronic myocardial ischemia, but also reduces the number of niche cells, which play a supportive role for CM function and entirely exacerbate the damage. The SBF-SEM is an innovative tool for 3D imaging and quantitative analysis of myocardial ultrastructure and functions.Methods & Figure 1Figure 2 and Figure 3

Interventricular abnormal septum displacement as contributory mechanism for impaired peak oxygen consumption and exercise cardiac output response in heart failure

Abstract Background In Heart failure (HF), the role of septum position and abnormal right ventricular (RV) geometry in determining exercise performance is unknown. Especially an abnormal interventricular septal (IVS) position is a well-known sign of RV pressure overload which affects left ventricular filling and output due to interdependence. Aim We postulated that an abnormal septum displacement during exercise and RV geometry alteration would affect peak exercise oxygen consumption (VO2) and cardiac output (CO) increase. Methods 17 HF patients with reduced ejection fraction (HFrEF) (mean age 71±12) underwent a cardiopulmonary exercise testing imaging (iCPET) with RV 3D-imaging analysis and were compared with a control population (mean age 67±14). 3D imaging of the RV chamber was examined off-line using the 4D RV TomTec software and obtained 3D mesh of the RV model using custom software to obtain the mean curvature value of IVS in 4 regions of: inflow tract (RVIT), outflow tract (RVOT), apical and body.We acquired measurements of curvature during end-diastole (ED) and at end systole (ES) phases and obtained a parametric curvature map. Results HF patients typically exhibited an abnormal of septal curve, with a more leftward configuration either at rest and under exercise (rest =−0.01±0.004 at ED, and −0.01±0.006 at ES; peak exercise= −0.01±0.004 at ED, and −0.01±0.004 at ES, Figure) compared to controls (rest=−0.02±0.003 at ED, and −0.02±0.005 at ES; peak exercise −0.02±0.008 at ED, and −0.03±0.006 at ES, Figure). Notably, the degree of the IVS curvature was found to linearly correlate with an impaired gas exchange performance as lower peak VO2 and a limited CO in HF (respectively r=0.62, p<0.001 and r=0.5, p<0.001 at ED during exercise; r=0.34, p<0.001 at ES during exercise, Figure). Conclusions In HF, the RV geometrical alterations reflected by an abnormal IVS displacement appear as novel contributory factors to a reduced exercise performance, impaired oxygen uptake and decreased CO.

Enhancing Grapevine Node Detection to Support Pruning Automation: Leveraging State-of-the-Art YOLO Detection Models for 2D Image Analysis

Automating pruning tasks entails overcoming several challenges, encompassing not only robotic manipulation but also environment perception and detection. To achieve efficient pruning, robotic systems must accurately identify the correct cutting points. A possible method to define these points is to choose the cutting location based on the number of nodes present on the targeted cane. For this purpose, in grapevine pruning, it is required to correctly identify the nodes present on the primary canes of the grapevines. In this paper, a novel method of node detection in grapevines is proposed with four distinct state-of-the-art versions of the YOLO detection model: YOLOv7, YOLOv8, YOLOv9 and YOLOv10. These models were trained on a public dataset with images containing artificial backgrounds and afterwards validated on different cultivars of grapevines from two distinct Portuguese viticulture regions with cluttered backgrounds. This allowed us to evaluate the robustness of the algorithms on the detection of nodes in diverse environments, compare the performance of the YOLO models used, as well as create a publicly available dataset of grapevines obtained in Portuguese vineyards for node detection. Overall, all used models were capable of achieving correct node detection in images of grapevines from the three distinct datasets. Considering the trade-off between accuracy and inference speed, the YOLOv7 model demonstrated to be the most robust in detecting nodes in 2D images of grapevines, achieving F1-Score values between 70% and 86.5% with inference times of around 89 ms for an input size of 1280 × 1280 px. Considering these results, this work contributes with an efficient approach for real-time node detection for further implementation on an autonomous robotic pruning system.

Real-Time 3D Imaging and Inhibition Analysis of Human Serum Amyloid A Aggregations Using Quantum Dots.

Serum amyloid A (SAA) is one of the most important precursor amyloid proteins discovered during the study of amyloidosis, but its underlying aggregation mechanism has not yet been well elucidated. Since SAA aggregation is a key step in the pathogenesis of AA amyloidosis, amyloid inhibitors can be used as a tool to study its pathogenesis. Previously, we reported a novel microliter-scale high-throughput screening (MSHTS) system for screening amyloid β (Aβ) aggregation inhibitors based on quantum dot (QD) fluorescence imaging technology. In this study, we report the aggregation of human SAA (hSAA) in phosphate-buffered saline, in which we successfully visualized hSAA aggregation by QD using fluorescence microscopy and confocal microscopy. Two-dimensional and three-dimensional image analyses showed that most aggregations were observed at 40 μM hSAA, which was the optimal aggregation concentration in vitro. The accuracy of this finding was verified by a Thioflavin T assay. The transmission electron microscopy results showed that QD uniformly bound to hSAA aggregation. hSAA aggregation inhibitory activity was also evaluated by rosmarinic acid (RA). The results showed that RA, which is a compound with high inhibitory activity against Aβ aggregation, also exhibited high inhibitory activity against 40 μM hSAA. These results indicate that the MSHTS system is an effective tool for visualizing hSAA aggregation and for screening highly active inhibitors.

Synergetic effect of hybrid surface treatment on the mechanical properties of adhesively bonded AA2024-T3 joints

ABSTRACT Improving the surface properties of AA2024-T3 aluminium alloys is crucial for achieving durable and sustainable joints. For this purpose, AA2024-T3 alloys were first sandblasted at 4, 5 and 6 bar pressures using aluminium oxide and glass sphere abrasives. Then, the sandblasted samples were subjected to the chemical surface treatment resulting in a hybrid surface process. The surface properties of sandblasted and hybrid surface-treated AA2024-T3 alloys were characterised by 3D image analysis, contact angle and roughness measurement. Additionally, single-lap joints were fabricated using adhesives with low and high viscosity. The effect of changes in the viscosity of adhesives and the surface properties of the AA2024-T3 on the mechanical properties of the joints has been analysed in detail. It was found that both sandblasting and hybrid surface treatments applied to AA2024-T3 alloys caused significant improvement in the mechanical properties of the joints. AA2024-T3 alloys, which were chemically surface treated after being sandblasted with glass spheres at 6 bar pressure, were joined using low and high viscosity adhesives, and the failure load of these joints increased by approximately 110% and 68%, respectively, compared to untreated samples. In the case of aluminium oxide usage, the increase rates were determined to be 93% and 43%.

Three-dimensional Stereoscopic Evaluation and Quantitative Gray Scale Assessment of the Nasolabial Fold and Its Application in Injection Therapy

With the growing demand for aesthetic enhancements, particularly filler injections for nasolabial folds (NLFs), there is a need for more objective and quantitative assessment methods beyond the traditional subjective scales like the Wrinkle Severity Rating Scale (WSRS). This study introduces a novel evaluation method combining 3-dimensional (3D) imaging and grayscale analysis to objectively assess NLF severity and treatment outcomes. A total of 100 healthy participants (86 females, 14 males; mean age 41.85±11.53 years) were assessed using WSRS scores and quantitative 3D measurements. The results demonstrated a significant positive correlation between WSRS scores and gray values (r=0.985, P < 0.05). Significant differences in gray values and 3D linear measurements were observed across WSRS score groups, with high intra-rater reliability confirmed by intraclass correlation coefficients. In addition, 54 patients (mean age 42.9±8.9 y) with moderate to severe NLFs received hyaluronic acid injections. One month post-treatment, WSRS scores significantly decreased (from 3.74±0.45 to 1.63±0.59, P < 0.05), and significant improvements were observed in several 3D measurements. The average gray value also decreased significantly (P < 0.05), indicating a reduction in wrinkle depth. These findings were supported by representative photos showing patient satisfaction and no severe complications. This study validates the use of 3D imaging and grayscale analysis as reliable tools for objectively assessing NLF severity and monitoring treatment efficacy, potentially improving the precision and reliability of aesthetic evaluations in clinical practice.

A novel approach to assess internal pore structure and quantify tortuosity of clogged pervious concrete through image analysis techniques

ABSTRACT The major objective of this research was to develop an innovative approach to characterise the pore structure of pervious concrete (PC) mixtures and quantify tortuosity in pre- and post-clogged conditions through 2D and 3D image analysis. Four different types of PC mixtures were prepared and subjected to computed tomography (CT) scanning to retrieve morphological information covering over 1900 images, followed by tortuosity quantification of each of 200 interconnected pore paths using vascular modelling toolkit (VMTK). The porosity of PC obtained from laboratory-based tests was 10–25% lower than image analyses. Further, along the depth of PC, a significant decrease in the number of pores in the top 60–80 mm was observed due to clogging, which revealed that maintenance be adopted only in the top 80 mm. Control and clogged PC with smaller sized aggregates and lower w/c ratio were more tortuous compared to higher w/c ratio while opposite for larger sized aggregates, ascribed to the occurrence of shorter interconnected pore paths comprising larger pore diameter, remarking that permeability will depend on both tortuosity and path diameter. CT scanning was used in predicting tortuosity and pore interconnectivity of PC through VMTK approach extensively used in medical science and seldom used in engineering.

Reliability analysis of WBCT-derived 3D models for comparing preoperative and postoperative alignment in total ankle arthroplasty

BackgroundTraditional imaging techniques for total ankle arthroplasty (TAA) evaluation are limited by rotational bias and bone superimposition, highlighting the necessity for more precise assessment methods. The advent of weight-bearing computed tomography (WBCT) generated 3D enhance the visualization of foot and ankle alignment, offering unmatched detail. This study aims to assess the accuracy of preoperative and postoperative measurements in TAA across all three planes using WBCT-generated 3D models. We hypothesize that these models can be reliably used to compare preoperative and postoperative alignment. MethodsFor 81 patients undergoing TAA, preoperative and postoperative WBCT models were created. Measurements included five coronal angles: Alpha, Tibiotalar Surface Angle (TSA), Talar Tilt Angle (TT), Salzmann's 20 degrees Angle (SA), and Talocalcaneal Angle (TCA); three sagittal angles: Beta, Gamma, and Tibiotalar Ratio (TTR); and one axial angle: The Posterior Talar Rotational Angle (PTARA). Two raters evaluated these before and after surgery in two separate sessions. The study then compared preoperative to postoperative measurements, calculating inter-rater and intra-rater reliability. ResultsSignificant changes were observed in three coronal angles (TSA, TT, and SA) and two sagittal angles (Beta and Gamma), with P-values of 0.2, 0.007, 0.019, <0.001, and <0.001, respectively. No significant changes were noted in Alpha, TCA, TTR, and PTARA, with P-values of 0.2, 0.9, 0.2, and 0.6, respectively. Intra-rater and inter-rater reliability scores ranged from 0.885 to 0.97, indicating good to excellent interclass correlation across all planes, both pre-and postoperatively. ConclusionWBCT-generated 3D modeling and image analysis software have enabled a detailed comparison between preoperative alignment and postoperative TAA positioning across coronal, sagittal, and axial planes, revealing significant adjustments in coronal and sagittal alignments. The high reliability and reproducibility of these measurements affirm their value in preoperative planning in improving the accuracy of surgical interventions. Level of evidenceLevel III of evidence

Volumetric Classification: Unveiling the True Extent of Rotator Cuff Tears

Rotator cuff injury diagnosis involves comprehensive clinical, physical, and imaging assessments, with MRI being pivotal for detecting and classifying these injuries. However, the absence of a universally accepted classification system necessitates a more precise approach, advocating for the use of three-dimensional (3D) modeling to better understand and categorize rotator cuff tears. This research was conducted as a prospective, single-institution study on 62 patients exhibiting full-thickness rotator cuff tears. Utilizing preoperative 1.5T MRI, the study aimed to create a more detailed classification system based on volumetric and surface area measurements. Advanced 3D modeling software was employed to transform MRI data into precise 3D representations, facilitating a more accurate analysis of the lesions. The study unveiled a novel classification system rooted in volumetric and surface area assessments, revealing significant discrepancies in the existing two-dimensional classifications. Approximately 45% of the cases demonstrated inconsistencies between traditional classifications and 3D measurements. Notably, medium-sized lesions were often overestimated, while small and large lesions were consistently underestimated in their severity. The volumetric and surface area-based classifications provided a more accurate depiction, highlighting the limitations of relying solely on coronal plane assessments in MRI. Comparative analysis confirmed the improved accuracy of the 3D method. The integration of 3D imaging and volumetric analysis offers novel advancement in diagnosing and classifying rotator cuff injuries. This study's findings challenge the conventional reliance on 2D MRI, proposing a more detailed and accurate classification system that enhances the precision of surgical planning and potentially improves patient outcomes. The incorporation of comprehensive 3D assessments into the diagnostic process represents a significant step forward in the orthopedic imaging field.

Evaluation of the Spike Diversity of Seven Hexaploid Wheat Species and an Artificial Amphidiploid Using a Quadrangle Model Obtained from 2D Images

The spike shape and morphometric characteristics are among the key characteristics of cultivated cereals, being associated with their productivity. These traits are often used for the plant taxonomy and authenticity of hexaploid wheat species. Manual measurement of spike characteristics is tedious and not precise. Recently, the authors of this study developed a method for wheat spike morphometry utilizing 2D image analysis. Here, this method is applied to study variations in spike size and shape for 190 plants of seven hexaploid (2n = 6x = 42) species and one artificial amphidiploid of wheat. Five manually estimated spike traits and 26 traits obtained from digital image analysis were analyzed. Image-based traits describe the characteristics of the base, center and apex of the spike and common parameters (circularity, roundness, perimeter, etc.). Estimates of similar traits by manual measurement and image analysis were shown to be highly correlated, suggesting the practical importance of digital spike phenotyping. The utility of spike traits for classification into types (spelt, normal and compact) and species or amphidiploid is shown. It is also demonstrated that the estimates obtained made it possible to identify the spike characteristics differing significantly between species or between accessions within the same species. The present work suggests the usefulness of wheat spike shape analysis using an approach based on characteristics obtained by digital image analysis.

Local mapping of root orientation traits by X-ray micro-CT and 3d image analysis: A study case on carrot seedlings grown in simulated vs real weightlessness.

Root phenotyping is particularly challenging because of complexity and inaccessibility of root apparatus. Orientation is one of the most important architectural traits of roots and its characterization is generally addressed using multiple approaches often based on overall measurements which are difficult to correlate to plant specific physiological aspects and its genetic features. Hence, a 3D image analysis approach, based on the recent method of Straumit, is proposed in this study to obtain a local mapping of root angles. Proposed method was applied here on radicles of carrot seedlings grown in real weightlessness on the International Space Station (ISS) and on Earth simulated weightlessness by clinorotation. A reference experiment in 1g static condition on Earth was also performed. Radicles were imaged by X-ray micro-CT and two novel root orientation traits were defined: the "root angle to sowing plane" (RASP) providing accurate angle distributions for each analysed radicle and the "root orientation changes" (ROC) number. The parameters of the RASP distributions and the ROC values did not exhibit any significant difference in orientation between radicles grown under clinorotation and on the ISS. Only a slight thickening in root corners was found in simulated vs real weightlessness. Such results showed that a simple uniaxial clinostat can be an affordable analog in experimental studies reckoning on weightless radicles growth. The proposed local orientation mapping approach can be extended also to different root systems providing a contribution in the challenging task of phenotyping complex and important plant structures such as roots.

Transforming Heart Failure Management: The Power of Strain Imaging, 3D Imaging, and Vortex Analysis in Echocardiography.

Heart failure (HF) remains a critical global health challenge, necessitating advancements in diagnostic and therapeutic strategies. This review explores the evolution of imaging technologies and their impact on HF management, focusing on three-dimensional echocardiography (3DE), myocardial strain imaging, and vortex dynamics imaging. Three-dimensional echocardiography enhances traditional echocardiography by providing more accurate assessments of cardiac structures, while myocardial strain imaging offers the early detection of subclinical myocardial dysfunction, crucial in conditions such as chemotherapy-induced cardiotoxicity and ischemic heart disease. Vortex dynamics imaging, a novel technique, provides insights into intracardiac flow patterns, aiding in the evaluation of left ventricular function, valve diseases, and congenital heart anomalies. The integration of these advanced imaging modalities into clinical practice facilitates personalized treatment strategies, enabling the earlier diagnosis and more precise monitoring of disease progression. The ongoing refinement of these imaging techniques holds promise for improving patient outcomes and advancing the field of precision medicine in HF care.

Exploring traditional and novel spike traits associated with seed retention in perennial ryegrass by integrating 2D and 3D image analysis

Exploring traditional and novel spike traits associated with seed retention in perennial ryegrass by integrating 2D and 3D image analysis

Timing of Laser Intervention on Facial Scars: A 3D Imaging and Scar Scale Analysis in a Retrospective Study.

Scars resulting from injuries or surgical procedures often present both physical and aesthetic challenges. Recent studies have indicated promising results in improving postoperative scar outcomes through the combined use of specific laser technologies. Nevertheless, there remains a crucial need for further exploration to ascertain the optimal timing for initiating such treatments. In this retrospective investigation, a cohort of 47 adult patients who did not require hospitalization was analyzed. These patients were divided into two distinct groups: Group A, which received intervention beginning 2 weeks after their respective operations, and Group B, which commenced intervention 4 weeks postoperation. Each group underwent a comprehensive treatment protocol consisting of five laser sessions. Initially, patients underwent three sessions of V-beam pulsed-dye laser (PDL) therapy followed by two sessions of ablative fractional CO2 laser (AFL) therapy. Evaluation of outcomes was conducted using advanced imaging techniques such as Antera 3D imaging, in conjunction with established scar assessment tools including the Vancouver Scar Scale (VSS) and the University of North Carolina "4P" Scar Scale (UNC4P). All patients successfully completed the full course of five treatment sessions, with no dropouts and no reported adverse events. Baseline cosmetic assessments of scars were equivalent across both groups. In Group A, the mean VSS scores decreased from 7.04 before treatment to 5.29 at 3 months posttreatment and further to 4.33 at 6 months posttreatment. Meanwhile, in Group B, scores decreased from 7.52 to 6.83 at 3 months and 6.17 at 6 months. There were no statistically significant differences in baseline VSS scores between the two groups (p = 0.34). At both the 3- and 6-month follow-up points, mean VSS scores were significantly lower in Group A compared to Group B (p < 0.05). Similar trends were observed in UNC4P scores. Statistically significant differences were noted across all time points (baseline, 3 months, and 6 months posttreatment) for both VSS and UNC4P scores (p < 0.05). Subset analysis revealed greater improvements in texture and depression volume at 3 months posttreatment, while improvements in depression area, depth, elevation variation, and area were more pronounced at the 6-month mark. Early intervention using the combined 595-nm pulse dye laser and CO2 ablative fractional laser (AFL) 2- and 4-weeks post-surgery proves to be an effective and safe method for improving scar outcomes, particularly for facial surgical scars when treatment starts 2 weeks after surgery. However, further research is needed to refine our understanding and address potential study limitations.

Dynamics of localized accelerated corrosion in reinforced concrete: Voids, corrosion products, and crack formation

A comprehensive 3D imaging analysis was conducted to investigate the corrosion process in reinforced concrete (RC) samples at various stages of the accelerated galvanostatic corrosion process. A state-of-the-art X-ray computed tomography (CT) scan technology was utilized to collect high-resolution 3D images of the specimen. The CT imaging was complemented and validated using a multi-faceted approach utilizing Scanning electron microscopy (SEM) and Raman spectrometry techniques. The distribution and evolution of voids, corrosion products, and cracks were investigated. A localized corrosion was observed that has some similarity to pitted corrosion but not as wide spread along the bar. The micro-scale imaging investigations revealed a gradual increase in the diameter of voids as the time of corrosion increased, especially after 8 days of accelerated corrosion process at which a significant decrease in small-sized voids (≤0.11 mm) accompanied by an increase in other void sizes was observed. At 10 and 12 days of accelerated corrosion process, the volume of larger-sized voids (≥0.35 mm) continued to increase, indicating a rapid corrosion progress and the formation of corrosion-induced cracks. In addition, a thorough X-ray CT analysis allowed us to differentiate between various categories of corrosion products. Notably, iron oxides and iron hydroxides were found to dominate the corrosion products. Understanding the composition and distribution of these corrosion products is essential as they play a significant role in the degradation and deterioration of the concrete structure. The coexistence of iron oxides and iron hydroxides in the region of corrosion products were confirmed using Raman spectroscopy analysis. The CT scan images captured the evolution of cracks at different time intervals, revealing a strong correlation between the initiation of cracks and the presence of corrosion products. Subsequently, during the cracking in the matrix, while the propagation of corrosion products within the matrix contributed to the expansion of cracks. SEM images and elemental analysis confirmed that these cracks were filled with corrosion products.

Fiber-type composition and 3D capillary analysis of the human splenius capitis muscle.

Despite the significance of neck muscles in musculoskeletal disorders, their microscopic anatomy remains poorly characterized. This study examined the splenius capitis muscle, focusing on its fiber-type composition, fiber size, and capillary network characteristics. For comparison and validation, the vastus lateralis muscle was also analyzed. Muscle samples from 13 young male subjects (mean age ± SD: 35.7 ± 8.6 years) were collected within 24-h post-mortem during autopsy. Myosin heavy chain (MyHC) isoform expression was characterized immunohistochemically in 10 μm sections, while the capillary network architecture was assessed in 100 μm sections. Immunofluorescence staining, confocal microscopy, and 3D image analysis were employed to quantify capillary tortuosity, anisotropy, branch density (Br dens), and the length of capillaries per muscle volume (LV), per muscle fiber length (LL), per fiber surface area (LS), and per fiber volume (LVf). Compared to the vastus lateralis muscle, the splenius capitis muscle had a higher percentage of type 1 fibers (51.2% vs 39.7%), fewer type 2a fibers (16.2% vs 31.4%), and smaller fiber diameters (35.5-40.9 μm vs 47-56.1 μm). It also displayed lower Br dens (P = 0.0069), higher anisotropy (P = 0.0004), and lower LL (P < 0.0001) but higher LVf (P = 0.0486). In the splenius capitis muscle, body mass index (BMI) negatively correlated with LV (P = 0.0155), LS (P = 0.0091), LVf (P = 0.0137), and anisotropy (P = 0.0425), and positively correlated with tortuosity (P = 0.0473), indicating a reduction in the capillary network. In the vastus lateralis muscle, only LV (P = 0.0161) decreased with high BMI. This study characterized the fiber-type composition, fiber size, and 3D capillary network of the splenius capitis muscle, establishing a baseline for investigations into pathological muscle alterations.

Weld seam object detection system based on the fusion of 2D images and 3D point clouds using interpretable neural networks

This study introduces a novel approach that addresses the limitations of existing methods by integrating 2D image processing with 3D point cloud analysis, enhanced by interpretable neural networks. Unlike traditional methods that rely on either 2D or 3D data alone, our approach leverages the complementary strengths of both data types to improve detection accuracy in environments adversely affected by welding spatter and smoke. Our system employs an improved Faster R-CNN model with a ResNet50 backbone for 2D image analysis, coupled with an innovative orthogonal plane intersection line extraction algorithm for 3D point cloud processing. By incorporating explainable components such as visualizable feature maps and a transparent region proposal network, we address the “black box” issue common in deep learning models.This architecture enables a more transparent decision-making process, providing technicians with necessary insights to understand and trust the system’s outputs. The Faster-RCNN structure is designed to break down the object detection process into distinct, understandable steps, from initial feature extraction to final bounding box refinement. This fusion of 2D-3D data analysis and interpretability not only improves detection performance but also sets a new standard for transparency and reliability in automated welding systems, facilitating wider adoption in industrial applications.

Application of Convolutional Neural Networks to time domain astrophysics. 2D image analysis of OGLE light curves

In recent years the amount of publicly available astronomical data has increased exponentially, with a remarkable example being large-scale multiepoch photometric surveys. This wealth of data poses challenges to the classical methodologies commonly employed in the study of variable objects. As a response, deep learning techniques are increasingly being explored to effectively classify, analyze, and interpret these large datasets. In this paper we use two-dimensional histograms to represent Optical Gravitational Lensing Experiment (OGLE) phasefolded light curves as images. We use a Convolutional Neural Network (CNN) to classify variable objects within eight different categories (from now on labels): Classical Cepheid (CEP), RR Lyrae (RR), Long Period Variable (LPV), Miras (M), Ellipsoidal Binary (ELL), Delta Scuti (DST), Eclipsing Binary (E), and spurious class with Incorrect Periods (Rndm). We set up different training sets to train the same CNN architecture in order to characterize the impact of the training. The training sets were built from the same source of labels but different filters and balancing techniques were applied. Namely: Undersampling (U), Data Augmentation (DA), and Batch Balancing (BB). The best performance was achieved with the BB approach and a training sample size of sim 370000 stars. Regarding computational performance, the image representation production rate is of sim 76 images per core per second, and the time to predict is sim 60$\ s $ per star. The accuracy of the classification improves from sim 92&lt;!PCT!&gt;, when based only on the CNN, to sim 98&lt;!PCT!&gt; when the results of the CNN are combined with the period and amplitude features in a two step approach. This methodology achieves comparable results with previous studies but with two main advantages: the identification of miscalculated periods and the improvement in computational time cost.