3D Measurement Research Articles

StrainNet-LD: Large Displacement digital image correlation based on deep learning and displacement-field decomposition

First, a displacement-field decomposition strategy is proposed that effectively enhances the accuracy of digital image correlation (DIC) algorithms under large displacement or deformation conditions. The implementation methodologies of displacement-field decomposition under both Lagrangian and Eulerian descriptions are derived and presented. Second, displacement-field decomposition is applied to speckle image correlation algorithms based on deep learning, resolving the contradiction between registration capability of large displacement and the accuracy. A CUDA-accelerated and residual learning-based strain fitting algorithm is proposed for high-precision and real-time strain computation. This method is named “StrainNet for Large Displacement” (StrainNet-LD). StrainNet-LD achieves a calculation speed of 768 × 768 × 6 fps (768 × 768 × 3 fps with strain calculated simultaneously) on GPU with displacement MAE < 0.05 pixels and strain MAE < 0.002ε. Finally, rubber tensile tests, 3D motion measurement of composite shells, and morphology reconstruction experiments of speckle-free composite repair structures are conducted to validate the algorithm's robustness and efficiency. Some vital discussions of displacement-field decomposition are given in the Appendixes. The code and dataset are available at https://github.com/GW-Wang-thu/StrainNet-LD.

Morphological changes in flatfoot: a 3D analysis using weight-bearing CT scans

BackgroundFlatfoot is a condition resulting from complex three-dimensional (3D) morphological changes. Most Previous studies have been constrained by using two-dimensional radiographs and non-weight-bearing conditions. The deformity in flatfoot is associated with the 3D morphology of the bone. These morphological changes affect the force line conduction of the hindfoot/midfoot/forefoot, leading to further morphological alterations. Given that a two-dimensional plane axis overlooks the 3D structural information, it is essential to measure the 3D model of the entire foot in conjunction with the definition under the standing position. This study aims to analyze the morphological changes in flatfoot using 3D measurements from weight-bearing CT (WBCT).MethodIn this retrospective comparative our CT database was searched between 4–2021 and 3–2022. Following inclusion criteria were used: Patients were required to exhibit clinical symptoms suggestive of flatfoot, including painful swelling of the medial plantar area or abnormal gait, corroborated by clinical examination and confirmatory radiological findings on CT or MRI. Healthy participants were required to be free of any foot diseases or conditions affecting lower limb movement. After applying the exclusion criteria (Flatfoot with other foot diseases), CT scans (mean age = 20.9375, SD = 16.1) confirmed eligible for further analysis. The distance, angle in sagittal/transverse/coronal planes, and volume of the two groups were compared on reconstructed 3D models using the t-test. Logistic regression was used to identify flatfoot risk factors, which were then analyzed using receiver operating characteristic curves and nomogram.ResultThe flatfoot group exhibited significantly lower values for calcaneofibular distance (p = 0.001), sagittal and transverse calcaneal inclination angle (p < 0.001), medial column height (p < 0.001), sagittal talonavicular coverage angle (p < 0.001), and sagittal (p < 0.001) and transverse (p = 0.015) Hibb angle. In contrast, the sagittal lateral talocalcaneal angle (p = 0.013), sagittal (p < 0.001) and transverse (p = 0.004) talocalcaneal angle, transverse talonavicular coverage angle (p < 0.001), coronal Hibb angle (p < 0.001), and sagittal (p < 0.001) and transverse (p = 0.001) Meary’s angle were significantly higher in the flatfoot group. The sagittal Hibb angle (B = − 0.379, OR = 0.684) and medial column height (B = − 0.990, OR = 0.372) were identified as significant risk factors for acquiring a flatfoot.ConclusionThe findings validate the 3D spatial position alterations in flatfoot. These include the abduction of the forefoot and prolapse of the first metatarsal proximal, the arch collapsed, subluxation of the talonavicular joint in the midfoot, adduction and valgus of the calcaneus, adduction and plantar ward movement of the talus in the hindfoot, along with the first metatarsal’s abduction and dorsiflexion in the forefoot.

3D displacement measurement using a single-camera and mesh deformation neural network

Computer vision-based methods for civil structure’s vibration displacement measurement have emerged as useful tools in the recent years. These methods offer several benefits including non-contact measurements, cost-effectiveness, and the ability to capture full-field displacement. Yet, there remain certain challenges. Measuring vibration displacement in 3D typically requires multiple cameras, adding complexity to camera configurations. Moreover, existing methods relied heavily on physical markers or natural key points. Placing physical markers on structures is often impractical, and natural key points are difficult to detect on structures with few distinct features or during rapid movements. Contrary to previous approaches, this paper presents a novel technique that uses a monocular camera for 3D displacement measurements. This technique obviates the need for physical markers or the reliance on natural key points, representing a significant advancement. Central to the method is a deep neural network designed to predict 3D mesh deformation directly from a single image input, combined with an initial 3D cube mesh input. A synthetic 3D dataset is generated to train the neural network. In the testing phase for real structures, advanced video segmentation method is employed to remove the background in order to enhance the prediction accuracy. The practical efficacy of this methodology is validated in a laboratory through a series of experimental tests on beam structures, demonstrating reliable results and application potentials.

Deep learning-based quantification of brain atrophy using 2D T1-weighted MRI for Alzheimer's disease classification.

Determining brain atrophy is crucial for the diagnosis of neurodegenerative diseases. Despite detailed brain atrophy assessments using three-dimensional (3D) T1-weighted magnetic resonance imaging, their practical utility is limited by cost and time. This study introduces deep learning algorithms for quantifying brain atrophy using a more accessible two-dimensional (2D) T1, aiming to achieve cost-effective differentiation of dementia of the Alzheimer's type (DAT) from cognitively unimpaired (CU), while maintaining or exceeding the performance obtained with T1-3D individuals and to accurately predict AD-specific atrophy similarity and atrophic changes [W-scores and Brain Age Index (BAI)]. Involving 924 participants (478 CU and 446 DAT), our deep learning models were trained on cerebrospinal fluid (CSF) volumes from 2D T1 images and compared with 3D T1 images. The performance of the models in differentiating DAT from CU was assessed using receiver operating characteristic analysis. Pearson's correlation analyses were used to evaluate the relations between 3D T1 and 2D T1 measurements of cortical thickness and CSF volumes, AD-specific atrophy similarity, W-scores, and BAIs. Our deep learning models demonstrated strong correlations between 2D and 3D T1-derived CSF volumes, with correlation coefficients r ranging from 0.805 to 0.971. The algorithms based on 2D T1 accurately distinguished DAT from CU with high accuracy (area under the curve values of 0.873), which were comparable to those of algorithms based on 3D T1. Algorithms based on 2D T1 image-derived CSF volumes showed high correlations in AD-specific atrophy similarity (r = 0.915), W-scores for brain atrophy (0.732 ≤ r ≤ 0.976), and BAIs (r = 0.821) compared with those based on 3D T1 images. Deep learning-based analysis of 2D T1 images is a feasible and accurate alternative for assessing brain atrophy, offering diagnostic precision comparable to that of 3D T1 imaging. This approach offers the advantage of the availability of T1-2D imaging, as well as reduced time and cost, while maintaining diagnostic precision comparable to T1-3D.

Individualized generation of women's prototype based on the classification of body shape

The traditional prototype method only relies on the basic dimensions of the human body such as bust girth and back length to make the prototype pattern, ignoring the differences of human body details dimensions. In order to improve the fit of garments, the research on intelligent generation technology of individualized garment patterns has become one of the hot spots in the field of clothing. Based on the 3D body scanning technology, we proposed a method for generating prototype pattern based on body shape classification. The proposed method takes as inputs the body parameters (width, thickness and angle), while the output of the method is one of the four pattern generation rules-"Y" shape, "V" shape, "I" shape, or "X" shape. 207 subjects were divided into four categories based on the body angle parameters, namely "Y" shape, "V" shape, "I" shape and "X" shape, and a shape recognition model was built. The pattern database is required to obtain pattern generation rules by 3D point cloud reconstruction and flattening. Combined with the human body shape parameters, the mathematical formulas of the feature points on the pattern landmarks are built. The accuracy of body shape recognition model is 97.4%. The error analysis of the predicted pattern parameters shows that 90% of the pattern feature points have a goodness of fit above 0.7. In the 5 main landmarks, the proportion of pattern within the absolute error range is more than 80%, indicating that the prediction effect of the pattern is good. This method can be applied to the process of automatic pattern generation system based 2D measurement to improve work efficiency.

Volumetric Analyses of Dysmorphic Maxillofacial Structures Using 3D Surface-Based Approaches: A Scoping Review.

Background/Objectives: Three-dimensional (3D) analysis of maxillofacial structures in dysmorphic patients offers clinical advantages over 2D analysis due to its high accuracy and precision in measuring many morphological parameters. Currently, no reliable gold standard exists for calculating 3D volumetric measurements of maxillofacial structures when captured by 3D surface imaging techniques. The aim of this scoping review is to provide an overview of the scientific literature related to 3D surface imaging methods used for volumetric analysis of the dysmorphic maxillofacial structures of patients affected by CL/P or other syndromes and to provide an update on the existing protocols, methods, and, when available, reference data. Methods: A total of 17 papers selected according to strict inclusion and exclusion criteria were reviewed for the qualitative analysis out of more than 4500 articles published between 2002 and 2024 that were retrieved from the main electronic scientific databases according to the PRISMA-ScR guidelines. A qualitative synthesis of the protocols used for the selection of the anatomical areas of interest and details on the methods used for the calculation of their volume was completed. Results: The results suggest a great degree of heterogeneity between the reviewed studies in all the aspects analysed (patient population, anatomical structure, area selection, and volume calculation), which prevents any chance of direct comparison between the reported volumetric data. Conclusions: Our qualitative analysis revealed dissimilarities in the procedures specified in the studies, highlighting the need to develop uniform methods and protocols and the need for comparative studies to verify the validity of methods in order to achieve high levels of scientific evidence, homogeneity of volumetric data, and clinical consensus on the methods to use for 3D volumetric surface-based analysis.

Evaluation of automatic cochlear dimension measurement using ALPACA: a comparative study

Background Cochlear dimension measurements are critical in diagnosing and managing congenital sensorineural hearing loss. Objectives To evaluate the feasibility and reliability of an automated landmark approach for measuring cochlear dimensions (A-, B- and H-values). Material and Methods Cochlear parameters from 100 patients were measured by MPR, manual three-dimensional and ALPACA. We assessed intra- and inter-observer reliability as well as inter-method reliability. Statistical analyses were conducted to detect differences between the right and left ears, as well as to assess the relevance of the values obtained using ALPACA. Results All A-, B-, and H-values measured by the various methods showed a high intra-observer reliability with intra-class correlation coefficients (ICC) ranging from 0.70 to 0.99, and values gained by ALPACA reaching the highest ICC. Inter-method reliability was at a good level with ICC ranging from 0.51 to 0.86. There were no significant differences between the right and left ears’ measured values. Obvious positive correlations existed among cochlear dimensions measured by ALPACA. Conclusions and Significance The ALPACA method can be used to measure cochlear dimensions. Values obtained by the method demonstrate high reliability and consistency with a significant reduction in intra-observer variability compared to results from conventional MPR and manual 3D measurements.

Research on Path Planning Technology of a Line Scanning Measurement Robot Based on the CAD Model

With the development of robotics and vision measurement technology, the use of robots with line laser scanners for 3D scanning and measurement of parts has become a mainstream trend in the field of industrial inspection. Traditional scanning and measuring robots mainly use the teach-in scanning method, which has unstable scanning quality and low scanning efficiency. In this paper, the adaptive sampling method for a free-form surface, which can realize the adaptive distribution of surface measurement points according to the curvature features of free-form surfaces, is proposed first. Then, integrated with the proposed adaptive sampling method, the automatic path planning method is proposed. This method consists of adaptive sampling, scanning attitude calculation based on a quaternion, scanning viewpoint planning based on viewable cones, and scan path generation based on bi-directional scanning. Based on the proposed automatic path planning method, the scanning and measuring robot can obtain complete 3D information of the surface to be measured with high measurement accuracy and efficiency. The performance index of the laser scanner can be fully reached.

Usefulness of Olfactory Bulb Measurement in 3D-FIESTA in Differentiating Parkinson Disease from Atypical Parkinsonism.

Parkinson disease is a prevalent disease, with olfactory dysfunction recognized as an early nonmotor manifestation. It is sometimes difficult to differentiate Parkinson disease from atypical parkinsonism using conventional MR imaging and motor symptoms. It is also known that olfactory loss occurs to a lesser extent or is absent in atypical parkinsonism. To the best of our knowledge, no study has examined olfactory bulb changes to differentiate Parkinson disease from atypical parkinsonism, even in an early diagnosis, and its association with conventional MR imaging findings. Hence, we aimed to assess the utility of olfactory bulb measurements in differentiating Parkinson disease from atypical parkinsonism even in the early stage. In this retrospective study, we enrolled 108 patients with Parkinson disease, 13 with corticobasal syndrome, 15 with multiple system atrophy, and 17 with progressive supranuclear palsy who developed parkinsonism. Thirty-nine age-matched healthy subjects served as controls. All subjects underwent conventional MR imaging and 3D FIESTA for olfactory bulb measurements using manual ROI quantification of the cross-sectional olfactory bulb area using the coronal plane. Bilateral olfactory bulb measurements were averaged. For group comparisons, we used the Welch t test, and we assessed diagnostic accuracy using receiver operating characteristic analysis. Patients with Parkinson disease had a mean olfactory bulb area of 4.2 (SD, 1.0 mm2), significantly smaller than in age-matched healthy subjects (6.6 [SD, 1.7 mm2], P < .001), and those with corticobasal syndrome (5.4 [SD, 1.2 mm2], P < .001), multiple system atrophy (6.5 [SD, 1.2 mm2], P < .001), and progressive supranuclear palsy (5.4 [SD, 1.2 mm2], P < .001). The receiver operating characteristic analysis for the olfactory bulb area measurements showed good diagnostic performance in differentiating Parkinson disease from atypical parkinsonism, with an area under the curve of 0.87, an optimal cutoff value of 5.1 mm2, and a false-positive rate of 18%. When we compared within 2 years of symptom onset, the olfactory bulb in Parkinson disease (4.2 [SD, 1.1 mm2]) remained significantly smaller than in atypical parkinsonism (versus corticobasal syndrome (6.1 [SD, 0.7 mm2]), P < .001; multiple system atrophy (6.3 [SD, 1.4 mm2]), P < .001; and progressive supranuclear palsy (5.2 [1.3 mm2], P = .003, respectively). 3D FIESTA-based olfactory bulb measurement holds promise for distinguishing Parkinson disease from atypical parkinsonism, especially in the early stage.

Large-scale-adaptive fringe projection 3D measurement.

Fringe projection profilometry (FPP) faces significant challenges regarding calibration difficulty and stitching error accumulation when operating across scenes ranging from tens to hundreds of meters. This Letter presents a calibration-free 3D measurement method by integrating a binocular vision of a FPP scanner with a wide field-of-view (FoV) vision that constructs global benchmarks to unify local 3D scanning and global 3D stitching, which is adaptable to arbitrarily large-scale scenes. A posterior global optimization model is then established to determine the reconstruction parameters and stitching poses simultaneously at each scanning node with adaptively distributed benchmarks. Consequently, the integrated vision measurement system not only eliminates the large-scale pre-calibration and stitching error accumulation but also overcomes system structural instability during moving measurement. With the proposed method, we achieved 3D measurements with an accuracy of 0.25 mm and a density of 0.5 mm for over 50-m-long scenes.

Large-depth-range 3D measurement based on optimized multi-focal projection system

High-speed and large-depth-range 3D measurement technology is of great importance in a variety of fields. Conventional binary defocusing methods can achieve high measurement speed, but their depth range is limited because the defocusing effect is insufficient near the focal plane in the ordinary projection system. To address this problem, we propose an optimized multi-focal projection system by introducing a cylindrical lens with optimal parameter configuration. Compared to traditional methods, the proposed method could overcome the constraint of defocusing region to obtain high-quality sinusoidal fringe patterns over large depth range without sacrificing the measurement speed. In this paper, the principle of multi-focal projection system and the related scheme of parameter configuration are presented in order to illustrate the role of the cylindrical lens and system parameters on modulating the distribution of defocusing kernel and removing the limitations of defocusing region. Experimental results show that the proposed multi-focal projection system with optimal parameter configuration increased the depth range from 150 mm to 725 mm over the conventional single-focal system, achieving greater depth range and better measurement results.

A multi-directional redundant 3D-LPT system for ship–flight–deck wind interactions

In the past years, volumetric velocimetry measurements with helium-filled soap bubbles as tracer particles have been introduced in wind tunnel experiments and performed at large-scale, enabling the study of complex body aerodynamics. A limiting factor is identified in the field of wind engineering, where the flow around ships is frequently investigated. Considering multiple wind directions, the optical access for illumination and 3D imaging rapidly erodes the measurement regions due to shadows and incomplete triangulation. This work formalizes the concepts of volumetric losses and camera redundancy, and examines the performance of multi-directional illumination and imaging for monolithic and partitioned modes. The work is corroborated by experiments around a representative ship model. The study shows that a redundant system of cameras yields the largest measurement volume when partitioned into subsystems. The 3D measurements employing two illumination directions and seven cameras, yield the time-averaged velocity field around the ship. Regions of flow separation and recirculation are revealed, as well as sets of counter-rotating vortices in several stations from the ship bow to the flight–deck. The unsteady regime at the flight–deck is examined by proper orthogonal decomposition, indicating that the technique is suited for the analysis of large-scale unsteady flow features.

Comparative Evaluation of the Performance of a Mobile Device Camera and a Full-Frame Mirrorless Camera in Close-Range Photogrammetry Applications.

The comparative evaluation of the performance of a mobile device camera and an affordable full-frame mirrorless camera in close-range photogrammetry applications involves assessing the capabilities of these two types of cameras in capturing images for 3D measurement purposes. In this study, experiments are conducted to compare the distortion levels, the accuracy performance, and the image quality of a mobile device camera against a full-frame mirrorless camera when used in close-range photogrammetry applications in various settings. Analytical methodologies and specialized digital tools are used to evaluate the results. In the end, generalized conclusions are drawn for using each technology in close-range photogrammetry applications.

Deep learning-based prediction of 3-dimensional silver contact shapes enabling improved quality control in solar cell metallization

The industrial metallization of Si solar cells predominantly relies on screen printing, with silver as the preferred electrode material. However, the design of commercial screens often leads to suboptimal silver usage and increased electrical resistance due to print-related inhomogeneities like mesh marks, constrictions and spreading. Real-time monitoring of quality parameters during production has thus become increasingly critical. Current inline optical quality control systems usually only include 2D visualizations of the printed layout, which limits their effectiveness in quality control. Options that allow 3D measurements are usually slow, expensive, and therefore not worth considering in most cases. This research focuses on the development of a model that can estimate the three-dimensional shape of printed contact fingers from a single 2D image without the need of additional hardware using deep learning. Furthermore, a workflow for the generation of training data, which involves the creation of image pairs from a 2D microscope and a 3D confocal laser scanning microscope (CLSM) to accurately represent solar cell fingers, is presented. After model training, the predicted height maps are compared with the ground truth height maps, and the robustness of the model with respect to a paste variation and screen parameter variation is examined. The results confirm the feasibility and reliability of deep learning-based 3D shape estimation, extending its applicability to new, previously unseen data from screen-printed contact fingers. With a structural similarity index (SSIM) score of 0.76, a strong correlation between the estimated and ground truth height maps is established. In summary, our deep learning-based approach for height map estimation offers an effective and reliable solution for fast inline detection and analysis of the cross-sectional area of the printed contact fingers.

3D printing and CBCT anatomical reproducibility assessment in forensic scenarios

IntroductionThe scientific community highlighted the relevance of 3D physical models since the beginning of the XXI century, complementary to three-dimensional(3D) digital volume by computer tomography, to support court discussions on medico-legal issues. The recreation of 3D evidence can be an important tool for investigators and experts, providing a better understanding of the causes and circumstances of the events involved in a crime. ObjectiveThe present study aims to assess the reproducibility of 3D printed and 3D tomographic volumes generated from mandibles following simulated forensic injuries, highlighting the recreation of crime tools. Material and methodsConcerning the study design presented, data collection was performed in three phases. Nine simulated injuries of forensic interest were selected (phase1) and all the mandibles were scanned tomographically, individually, by Cone Beam Computed Tomography CBCT (phase 2). Then, in phase 3, the DICOM images were used for 3D printing with the Ender 3® printer by the Fused Deposition Modeling (FDM) technique. The data analysis followed two procedures: the comparison between the artificial mandible and 3D tomographic volume (AT) and the comparison between the artificial mandible and 3D printed volume, or the copy (AC). Data were analyzed using T-Student and ICC tests and presented in Bland-Altman plots. ConclusionThe analogic technique applied in 3D printed volume, when compared with computerized technique, using 3D digital images and measurement, showed to be accurate and reproducible. Further studies are needed in search of standardization for three-dimensional measurements in digitized and printed volumes.

Effect of cross-platform variations on transthoracic echocardiography measurements and clinical diagnosis.

Accurate cardiac chamber quantification is essential for clinical decisions and ideally should be consistent across different echocardiography systems. This study evaluates variations between the Philips EPIQ CVx (version 9.0.3) and Canon Aplio i900 (version 7.0) in measuring cardiac volumes, ventricular function, and valve structures. In this gender-balanced, single-centre study, 40 healthy volunteers (20 females and 20 males) aged 40 years and older (mean age 56.75 ± 11.57 years) were scanned alternately with both systems by the same sonographer using identical settings for both 2D and 4D acquisitions. We compared left ventricular (LV) and right ventricular (RV) volumes using paired t-tests, with significance set at P < 0.05. Correlation and Bland-Altman plots were used for quantities showing significant differences. Two board-certified cardiologists evaluated valve anatomy for each platform. The results showed no significant differences in LV end-systolic volume and LV ejection fraction between platforms. However, LV end-diastolic volume (LVEDV) differed significantly (biplane: P = 0.018; 4D: P = 0.028). Right ventricular (RV) measurements in 4D showed no significant differences, but there were notable disparities in 2D and 4D volumes within each platform (P < 0.01). Significant differences were also found in the LV systolic dyssynchrony index (P = 0.03), LV longitudinal strain (P = 0.04), LV twist (P = 0.004), and LV torsion (P = 0.005). Valve structure assessments varied, with more abnormalities noted on the Philips platform. Although LV and RV volumetric measurements are generally comparable, significant differences in LVEDV, LV strain metrics, and 2D vs. 4D measurements exist. These variations should be considered when using different platforms for patient follow-ups.

Feasibility, Accuracy, and Reproducibility of Aortic Valve Sizing for Transcatheter Aortic Valve Implantation Using Virtual Reality.

Detailed visualization and precise measurements of aortic valve dimensions are critical for the success of transcatheter aortic valve implantation and for the prevention of complications. Currently, multislice computed tomography is the gold standard for assessment of the aortic annulus and surrounding structures to determine the prosthesis size. New technologies such as virtual reality (VR) not only enable 3-dimensional (3D) visualization with the potential to improve understanding of anatomy and pathology but also allow measurements in 3D. This study aims to investigate the feasibility, accuracy, and reproducibility of VR for the visualization of the aortic valve, the surrounding structures, and its role in preprocedural sizing for transcatheter aortic valve implantation. Based on the preprocedural multislice computed tomography data, 3mensio measurements and 3D visualizations and measurements using VR software were performed retrospectively on 60 consecutive patients who underwent transcatheter aortic valve implantation at our heart center. There were no significant differences but strong correlations between the VR measurements compared with those performed with the 3mensio software. Furthermore, excellent or good intra- and interobserver reliability could be demonstrated for all values. In a structured questionnaire, users reported that VR simplified anatomical understanding, improved 3D comprehension of adjacent structures, and was associated with very good self-perceived depth perception. The use of VR for preprocedural transcatheter aortic valve implantation sizing is feasible and has precise and reproducible measurements. In addition, 3D visualization improves anatomical understanding and orientation. To evaluate the potential benefits of 3D visualization for planning further cardiovascular interventions, research in this field is needed.

Comprehensive stereotactic radiosurgery platform characterization: A novel end-to-end approach with anthropomorphic 3D dosimetry.

Stereotactic radiosurgery (SRS) is a widely employed strategy for intracranial metastases, utilizing linear accelerators and volumetric modulated arc therapy (VMAT). Ensuring precise linear accelerator performance is crucial, given the small planning target volume (PTV) margins. Rapid dose falloff is vital to minimize brain radiation necrosis. Despite advances in SRS planning, tools for end-to-end testing of SRS treatments are lacking, hindering confidence in the procedure. This study introduces a novel end-to-end three-dimensional (3D) anthropomorphic dosimetry system for characterization of a radiosurgery platform, aiming to measure planning metrics, dose gradient index (DGI), brain volumes receiving at least 10 and 12Gy (V10, V12), as well as assess delivery uncertainties in multitarget treatments. The study also compares metrics from benchmark plans to enhance understanding and confidence in SRS treatments. The developed anthropomorphic 3D dosimetry system includes a modified Stereotactic End-to-End Verification (STEEV) phantom with a customized insert integrating 3D dosimeters and a fiber optic CT scanner. Labview and MATLAB programs handle optical scanning, image preprocessing, and dosimetric analysis. SlicerRT is used for 3D dose visualization and analysis. A film stack insert was used to validate the 3D dosimeter measurements at specific slices. Benchmark plans were developed and measured to investigate off-axis errors, dose spillage, small field dosimetry, and multi-target delivery. The accuracy of the developed 3D dosimetry system was rigorously assessed using radiochromic films. Two two-dimensional (2D) dose planes, extracted from the 3D dose distribution, were compared with film measurements, resulting in high passing rates of 99.9% and 99.6% in gamma tests. The mean relative dose difference between film and 3D dosimeter measurements was -1%, with a standard deviation of 2.2%, well within dosimeter uncertainties. In the first module, evaluating single-isocenter multitarget treatments, a 1.5mm dose distribution shift was observed when targets were 7cm off-axis. This shift was attributed to machine mechanical errors and image-guided system uncertainties, indicating potential limitations in conventional gamma tests. The second module investigated discrepancies in intermediate-to-low dose spillage, revealing higher measured doses in the connecting region between closely positioned targets. This discrepancy was linked to uncertainties in treatment planning system (TPS) modeling of out-of-field dose and multileaf collimator (MLC) characteristics, resulting in lower DGI values and higher V10 and V12 values compared to TPS calculations. In the third module, irradiating multiple targets showed consistent V10 and V12 values within 1 cm3 agreement with dose calculations. However, lower DGI values from measurements compared to calculations suggested intricacies in the treatment process. Conducting vital end-to-end testing demonstrated the anthropomorphic 3D dosimetry system's capacity to assess overall treatment uncertainty, offering a valuable tool for enhancing treatment accuracy in radiosurgery platforms. The study introduces a novel anthropomorphic 3D dosimetry system for end-to-end testing of a radiosurgery platform. The system effectively measures plan quality metrics, captures mechanical errors, and visualizes dose discrepancies in 3D space. The comprehensive evaluation capability enhances confidence in the commissioning and verification process, ensuring patient safety. The system is recommended for commissioning new radiosurgery platforms and remote auditing of existing programs.

Characterization of structure and mixing in nanoparticle hetero-aggregates using convolutional neural networks: 3D-reconstruction versus 2D-projection

Structural and chemical characterization of nanomaterials provides important information for understanding their functional properties. Nanomaterials with characteristic structure sizes in the nanometer range can be characterized by scanning transmission electron microscopy (STEM). In conventional STEM, two-dimensional (2D) projection images of the samples are acquired, information about the third dimension is lost. This drawback can be overcome by STEM tomography, where the three-dimensional (3D) structure is reconstructed from a series of projection images acquired using various projection directions. However, 3D measurements are expensive with respect to acquisition and evaluation time. Furthermore, the method is hardly applicable to beam-sensitive materials, i.e. samples that degrade under the electron beam. For this reason, it is desirable to know whether sufficient information on structural and chemical information can be extracted from 2D-projection measurements. In the present work, a comparison between 3D-reconstruction and 2D-projection characterization of structure and mixing in nanoparticle hetero-aggregates is provided. To this end, convolutional neural networks are trained in 2D and 3D to extract particle positions and material types from the simulated or experimental measurement. Results are used to evaluate structure, particle size distributions, hetero-aggregate compositions and mixing of particles quantitatively and to find an answer to the question, whether an expensive 3D characterization is required for this material system for future characterizations.

An agent-based method to estimate 3D cell migration trajectories from 2D measurements: Quantifying and comparing T vs CAR-T 3D cell migration

Background and objective:Immune cell migration is one of the key features that enable immune cells to find invading pathogens, control tissue damage, and eliminate primary developing tumors. Chimeric antigen receptor (CAR) T-cell therapy is a novel strategy in the battle against various cancers. It has been successful in treating hematological tumors, yet it still faces many challenges in the case of solid tumors. In this work, we evaluate the three-dimensional (3D) migration capacity of T and CAR-T cells within dense collagen-based hydrogels. Quantifying three-dimensional (3D) cell migration requires microscopy techniques that may not be readily accessible. Thus, we introduce a straightforward mathematical model designed to infer 3D trajectories of cells from two-dimensional (2D) cell trajectories. Methods:We develop a 3D agent-based model (ABM) that simulates the temporal changes in the direction of migration with an inverse transform sampling method. Then, we propose an optimization procedure to accurately orient cell migration over time to reproduce cell migration from 2D experimental cell trajectories. With this model, we simulate cell migration assays of T and CAR-T cells in microfluidic devices conducted under hydrogels with different concentrations of type I collagen and validate our 3D cell migration predictions with light-sheet microscopy. Results:Our findings indicate that CAR-T cell migration is more sensitive to collagen concentration increases than T cells, resulting in a more pronounced reduction in their invasiveness. Moreover, our computational model reveals significant differences in 3D movement patterns between T and CAR-T cells. T cells exhibit migratory behavior in 3D whereas that CAR-T cells predominantly move within the XY plane, with limited movement in the Z direction. However, upon the introduction of a CXCL12 chemical gradient, CAR-T cells present migration patterns that closely resemble those of T cells. Conclusions:This framework demonstrates that 2D projections of 3D trajectories may not accurately represent real migration patterns. Moreover, it offers a tool to estimate 3D migration patterns from 2D experimental data, which can be easily obtained with automatic quantification algorithms. This approach helps reduce the need for sophisticated and expensive microscopy equipment required in laboratories, as well as the computational burden involved in producing and analyzing 3D experimental data.