Set Of Scans Research Articles

Adversarial Consistency for Single Domain Generalization in Medical Image Segmentation.

An organ segmentation method that can generalize to unseen contrasts and scanner settings can significantly reduce the need for retraining of deep learning models. Domain Generalization (DG) aims to achieve this goal. However, most DG methods for segmentation require training data from multiple domains during training. We propose a novel adversarial domain generalization method for organ segmentation trained on data from a single domain. We synthesize the new domains via learning an adversarial domain synthesizer (ADS) and presume that the synthetic domains cover a large enough area of plausible distributions so that unseen domains can be interpolated from synthetic domains. We propose a mutual information regularizer to enforce the semantic consistency between images from the synthetic domains, which can be estimated by patch-level contrastive learning. We evaluate our method for various organ segmentation for unseen modalities, scanning protocols, and scanner sites.

Development of a storage phosphor imaging system for proton pencil beam spot profile determination.

Accurate two-dimensional (2D) profile measurements at submillimeter precision are necessary for proton beam commissioning and periodic quality assurance (QA) purposes and are currently performed at our institution with a commercial scintillation detector (Lynx PT) with limited means for independent checks. The purpose of this work was to create an independent dosimetry system consisting of an in-house optical scanner and a BaFBrI:Eu2+ storage phosphor dosimeter by: (a) determining the optimal settings for the optical scanner, (b) measuring 2D proton spot profiles with the storage phosphors, and (c) comparing them to similar measurements using a commercial scintillation detector. An in-house 2D laboratory optical scanner was constructed and spatially calibrated for accurate 2D photostimulated luminescence (PSL) dosimetry. Square 5×5cm2 BaFBrI:Eu2+ dosimeter samples were uniformly irradiated with line scans performed to determine the physical and electronic scanner settings resulting in the highest signal-to-noise ratios (SNR) at a sub-millimeter spatial resolution. The resultant spatial resolution of the scanner was then quantitatively assessed by measuring (a) line pairs on a standard X-ray lead bar phantom and (b) modulation transfer functions. Following this, 2D proton spot profiles from a Mevion S250i Hyperscan proton unit were obtained at 1, 10, 20, 30, 40, and 50 monitor unit (MU) settings at maximum energy (E0 =227.1MeV) and compared to baseline profiles from a commercial scintillation detector, where 1MU is calibrated to deliver 1Gy absolute proton dose-to-water under reference conditions, that is, 41×41 proton spots uniformly spaced by 0.25cm within a 10×10cm2 square field size at maximum energy (227.1MeV) in water at depth of 5cm at isocenter. The dosimetric system's sensitivities to (a) ±1mm positional shifts and (b) ±0.3mm beam lateral spread changes were quantitatively evaluated through a Gaussian fitting of the crossline and inline plots of the respective artificially shifted beam profiles. The physical scanner settings of (a) Δτ=27ms time interval between data samples, (b) vx =1.235cm/s scanning speed, (c) 1% laser transmission (0.02mW power) and (d) (Δx, Δy)=(0.33, 0.50mm) pixel sizes with electronic settings of (a) 300 microseconds time constant, (b) normal dynamic reserve, (c) 24dB/oct low pass filter slope, and (d) 160Hz chopping frequency resulted in the highest SNR while maintaining sub-millimeter spatial resolution. The BaFBr0.85 I0.15 :Eu2+ storage phosphor dosimeters were linear from 1 to 50MU and their profiles did not saturate up to 150MU. The scanner was able to detect lateral displacements of ±1mm in both the crossline and inline directions and ±0.3mm beam spread changes that were artificially introduced by varying the incident proton energy. Specific to our proton unit, proton energy changes of ±1MeV can also be detected indirectly via beam spread measurements. Our combined dosimetric system including an in-house laboratory optical scanner and reusable BaFBr0.85 I0.15 :Eu2+ storage phosphors demonstrated a sufficient spatial resolution and dosimetric accuracy to support its use as an independent proton spot measurement dosimeter system. Its wide dynamic range allows for other versatile applications such as proton halo measurements.

Implication of applying iterative reconstruction on Low Contrast Detectability in CT brain examination

Iterative reconstruction (IR) of CT images makes it possible to reduce image noise while reducing radiation exposure. There are several impacts of IR techniques on image quality that were not definable when applied to CT head examination. In this study, we aim to evaluate the signal to noise ratio (SNR) and contrast to noise ratio (CNR) of CT brain examination as image quality performance and to establish indicators for image quality as optimization references. Thirty patients, aged ranged from 18 to 65 years old (55.50 ± 21.45) and had CT brain examinations within the period between June 2019 and September 2019 were recruited. The CT examinations were performed using a 128-slices CT-Scanner (Philips Brilliance iCT, Netherland) with standard scanning acquisition protocols of CT brain examinations. Patients were scanned in the helical mode with different scanner settings for the brain and reconstructed using Filtered Back Projection (FBP) and different levels of IR algorithm (iDose4 level 1 to 5). Objective image quality analysis was performed with a SNR and CNR by drawing region of interest (ROI) on the lens, pons, frontal lobe and cerebellum. Objective measurement shows an overall increase in SNR and slightly increase the CNR at a higher iDose4 level. Different levels of iDose4 affect the quality of an objective image in a CT brain exam. IR has great potential to improve CT performance: the image quality of the brainstem and cerebellum are improved, and the dose exposure can be lowered in the future through a prospective study.

Challenges in Building of Deep Learning Models for Glioblastoma Segmentation: Evidence from Clinical Data.

In this article, we compare the performance of a state-of-the-art segmentation network (UNet) on two different glioblastoma (GB) segmentation datasets. Our experiments show that the same training procedure yields almost twice as bad results on the retrospective clinical data compared to the BraTS challenge data (in terms of Dice score). We discuss possible reasons for such an outcome, including inter-rater variability and high variability in magnetic resonance imaging (MRI) scanners and scanner settings. The high performance of segmentation models, demonstrated on preselected imaging data, does not bring the community closer to using these algorithms in clinical settings. We believe that a clinically applicable deep learning architecture requires a shift from unified datasets to heterogeneous data.

Quantitative Muscle Ultrasonography Using 2D Textural Analysis: A Novel Approach to Assess Skeletal Muscle Structure and Quality in Chronic Kidney Disease.

Chronic kidney disease (CKD) is characterized by progressive reductions in skeletal muscle function and size. The concept of muscle quality is increasingly being used to assess muscle health, although the best means of assessment remains unidentified. The use of muscle echogenicity is limited by an inability to be compared across devices. Gray level of co-occurrence matrix (GLCM), a form of image texture analysis, may provide a measure of muscle quality, robust to scanner settings. This study aimed to identify GLCM values from skeletal muscle images in CKD and investigate their association with physical performance and strength (a surrogate of muscle function). Transverse images of the rectus femoris muscle were obtained using B-mode 2D ultrasound imaging. Texture analysis (GLCM) was performed using ImageJ. Five different GLCM features were quantified: energy or angular second moment (ASM), entropy, homogeneity, or inverse difference moment (IDM), correlation, and contrast. Physical function and strength were assessed using tests of handgrip strength, sit to stand-60, gait speed, incremental shuttle walk test, and timed up-and-go. Correlation coefficients between GLCM indices were compared to each objective functional measure. A total of 90 CKD patients (age 64.6 (10.9) years, 44% male, eGFR 33.8 (15.7) mL/minutes/1.73 m2) were included. Better muscle function was largely associated with those values suggestive of greater image texture homogeneity (i.e., greater ASM, correlation, and IDM, lower entropy and contrast). Entropy showed the greatest association across all the functional assessments (r = −.177). All GLCM parameters, a form of higher-order texture analysis, were associated with muscle function, although the largest association as seen with image entropy. Image homogeneity likely indicates lower muscle infiltration of fat and fibrosis. Texture analysis may provide a novel indicator of muscle quality that is robust to changes in scanner settings. Further research is needed to substantiate our findings.

Gaussian smoothing and modified histogram normalization methods to improve neural-biomarker interpretations for dyslexia classification mechanism.

Achieving biologically interpretable neural-biomarkers and features from neuroimaging datasets is a challenging task in an MRI-based dyslexia study. This challenge becomes more pronounced when the needed MRI datasets are collected from multiple heterogeneous sources with inconsistent scanner settings. This study presents a method of improving the biological interpretation of dyslexia’s neural-biomarkers from MRI datasets sourced from publicly available open databases. The proposed system utilized a modified histogram normalization (MHN) method to improve dyslexia neural-biomarker interpretations by mapping the pixels’ intensities of low-quality input neuroimages to range between the low-intensity region of interest (ROIlow) and high-intensity region of interest (ROIhigh) of the high-quality image. This was achieved after initial image smoothing using the Gaussian filter method with an isotropic kernel of size 4mm. The performance of the proposed smoothing and normalization methods was evaluated based on three image post-processing experiments: ROI segmentation, gray matter (GM) tissues volume estimations, and deep learning (DL) classifications using Computational Anatomy Toolbox (CAT12) and pre-trained models in a MATLAB working environment. The three experiments were preceded by some pre-processing tasks such as image resizing, labelling, patching, and non-rigid registration. Our results showed that the best smoothing was achieved at a scale value, σ = 1.25 with a 0.9% increment in the peak-signal-to-noise ratio (PSNR). Results from the three image post-processing experiments confirmed the efficacy of the proposed methods. Evidence emanating from our analysis showed that using the proposed MHN and Gaussian smoothing methods can improve comparability of image features and neural-biomarkers of dyslexia with a statistically significantly high disc similarity coefficient (DSC) index, low mean square error (MSE), and improved tissue volume estimations. After 10 repeated 10-fold cross-validation, the highest accuracy achieved by DL models is 94.7% at a 95% confidence interval (CI) level. Finally, our finding confirmed that the proposed MHN method significantly outperformed the normalization method of the state-of-the-art histogram matching.

First experience of efficacy and radiation exposure in 320-detector row CT fluoroscopy-guided interventions.

We investigated the efficacy and exposure to radiation in 320-detector row computed tomography fluoroscopy-guided (CTF-guided) interventions. We analysed 231 320-detector row CTF-guided interventions (207 patients over 2 years and 6 months) in terms of technical success rates, clinical success rates, complications, scanner settings, overall radiation doses (dose-length product, mGy*cm), patient doses of peri-interventional CT series, and interventional CT (including CTF), as a retrospective cohort study. The relationships between patient radiation dose and interventional factors were assessed using multivariate analysis. Overall technical success rate was 98.7% (228/231). The technical success rates of biopsies, drainages, and aspirations were 98.7% (154/156), 98.5% (66/67), and 100% (8/8), respectively. The clinical success rate of biopsies was 93.5% (146/156). All three major complications occurred in chest biopsies. The median total radiation dose was 522.4 (393.4-819.8) mGy*cm. Of the total radiation dose, 87% was applied during the pre- and post-interventional CT series. Post-interventional CT accounted for 24.4% of the total radiation dose. Only 11.4% of the dose was applied by CTF-guided intervention. Multilinear regression demonstrated that male sex, body mass index, drainage, intervention time, and helical scan as post-interventional CT were significantly associated with higher dose. The 320-detector row CTF interventions achieved a high success rate. Dose reduction in post-interventional CT provides patient dose reduction without decreasing the technical success rates. This is the first study on the relationship between various interventional outcomes and patient exposure to radiation in 320-detector row CTF-guided interventions, suggesting a new perspective on dose reduction.

Dual-Energy Computed Tomography for Detection and Characterization of Monosodium Urate, Calcium Pyrophosphate, and Hydroxyapatite: A Phantom Study on Diagnostic Performance.

The aim of this study was to determine the diagnostic performance of dual-energy computed tomography (DECT) to detect and distinguish crystal deposits in a phantom. The primary objective was to determine the cutoff DECT ratio and the cross-sectional area (CSA) of a crystal deposit necessary to differentiate monosodium urate (MSU), calcium pyrophosphate (CPP), and calcium hydroxyapatite (HA) using DECT. Our secondary objective was to determine the concentration for limit of detection for MSU, CPP, and HA crystal deposits. Exploratory objectives included the comparison between 2 generations of DECT scanners from the same manufacturer as well as different scanner settings. We used a cylindrical soft tissue phantom with synthetic MSU, CPP, and HA crystals suspended in resin. Crystal suspension concentration increased with similar attenuation between MSU, CPP, and HA in conventional CT. The phantom was scanned on 2 dual-source DECT scanners, at 2 dose levels and all available tube voltage combinations. Both scanners had a tin (Sn) filter at the high-energy spectra. Dual-energy CT ratios were calculated for a given tube voltage combination by dividing linear regression lines of CT numbers against concentration. Dual-energy CT ratios were compared using an analysis of covariance. Receiver operating characteristic curves and corresponding areas under the curve (AUCs) were calculated for individual crystal suspension comparisons (HA vs CPP, MSU vs CPP, and MSU vs HA). At standard clinical scan settings with 8 mGy and 80/Sn150 kV, the DECT ratios were as follows: CPP, 2.02 (95% confidence interval [CI], 1.98-2.07); HA, 2.00 (95% CI, 1.96-2.05); and MSU, 1.09 (95% CI, 1.06-1.11). Ratios varied numerically depending on the scanner and tube voltage combination. Monosodium urate crystal DECT ratios were significantly different from HA and CPP (P < 0.001), whereas DECT ratios for HA and CPP crystals did not differ significantly (P = 0.99). The differentiation of MSU crystals from both calcium crystals (HA and CPP) was excellent with an AUC of 1.00 (95% CI, 1.00-1.00) and an optimal cutoff DECT ratio of 1.43:1.40 depending on the scanner. In addition, differentiation of MSU and calcium-containing crystals (HA and CPP) required a CSA of minimum 4 pixels of crystal at standard clinical scan conditions. In contrast, differentiation between CPP and HA crystals was moderate with AUCs ranging from 0.66 (95% CI, 0.52-0.80) to 0.80 (95% CI, 0.69-0.91) and an optimal cutoff DECT ratio of 2.02:2.06 depending on the scanner. Furthermore, differentiation between CPP and HA crystals required a CSA of minimum 87 pixels of crystal at standard clinical scan conditions, corresponding to a region of interest of 3.7 mm diameter. When scanning at highest possible spectral separation and maximum dose of 50 mGy, the limit of detection for crystals within a region of interest of 50 pixels was 14 mg/cm3 for MSU and 2 mg/cm3 for both CPP and HA. This phantom study shows that DECT can be used to detect MSU, CPP, and HA crystal deposits. Differentiation of CPP and HA was not possible in crystals deposits less than 3.7 mm in diameter, but MSU could accurately be differentiated from CPP and HA crystal deposits at standard clinical scan conditions.

PWD-3DNet: A Deep Learning-Based Fully-Automated Segmentation of Multiple Structures on Temporal Bone CT Scans

The temporal bone is a part of the lateral skull surface that contains organs responsible for hearing and balance. Mastering surgery of the temporal bone is challenging because of this complex and microscopic three-dimensional anatomy. Segmentation of intra-temporal anatomy based on computed tomography (CT) images is necessary for applications such as surgical training and rehearsal, amongst others. However, temporal bone segmentation is challenging due to the similar intensities and complicated anatomical relationships among critical structures, undetectable small structures on standard clinical CT, and the amount of time required for manual segmentation. This paper describes a single multi-class deep learning-based pipeline as the first fully automated algorithm for segmenting multiple temporal bone structures from CT volumes, including the sigmoid sinus, facial nerve, inner ear, malleus, incus, stapes, internal carotid artery and internal auditory canal. The proposed fully convolutional network, PWD-3DNet, is a patch-wise densely connected (PWD) three-dimensional (3D) network. The accuracy and speed of the proposed algorithm was shown to surpass current manual and semi-automated segmentation techniques. The experimental results yielded significantly high Dice similarity scores and low Hausdorff distances for all temporal bone structures with an average of 86% and 0.755 millimeter (mm), respectively. We illustrated that overlapping in the inference sub-volumes improves the segmentation performance. Moreover, we proposed augmentation layers by using samples with various transformations and image artefacts to increase the robustness of PWD-3DNet against image acquisition protocols, such as smoothing caused by soft tissue scanner settings and larger voxel sizes used for radiation reduction. The proposed algorithm was tested on low-resolution CTs acquired by another center with different scanner parameters than the ones used to create the algorithm and shows potential for application beyond the particular training data used in the study.

Speckle Tracking-Derived Longitudinal Strain: Validation and Influence of Scanner Settings

Speckle tracking-based strain analysis is an evolving technique for the assessment of left ventricular function. We evaluated the influence of machine settings on global longitudinal peak systolic strain (GLPSS) values in an everyday patient population (n = 35). In each patient, the four-chamber view was recorded multiple times with different machine parameters. Ejection fraction ranged between 10% and 76% and correlated well with GLPSS (r = –0.778). GLPSS was not altered systematically by modifications of gain and frame rate. Conversely, higher transducer frequencies (mean effect: 1.102%/MHz, p < 0.001, and 0.662%/MHz, p = 0.033, for harmonic and fundamental imaging frequencies, respectively) and lower sector depth (mean effect: –0.156%/cm, p < 0.001) were associated with a slight—but statistically significant—reduction in absolute GLPSS values. Intra- and inter-observer variability exhibited satisfactory repeatability. GLPSS analysis proved to be reproducible and robust in our patient cohort if common settings for adult echocardiography were applied.

Pediatric contrast-enhanced ultrasound: optimization of techniques and dosing.

When performing contrast-enhanced ultrasound (CEUS), ultrasound (US) scanner settings, examination technique, and contrast agent dose and administration must be optimized to ensure that high-quality, diagnostic and reproducible images are acquired for qualitative and quantitative interpretations. When carrying out CEUS in children, examination settings should be tailored to their body size and specific indications, similar to B-mode US. This review article details the basic background knowledge that is needed to perform CEUS optimally in children, including considerations related to US scanner settings and US contrast agent dose selection and administration techniques.

Accuracy of CAD/CAM Digital Impressions with Different Intraoral Scanner Parameters.

The advancement of intraoral scanners has allowed for more efficient workflow in the dental clinical setting. However, limited data exist regarding the accuracy of the digital impressions produced with various scanner settings and scanning approaches. The purpose of this in vitro study was to compare the accuracy of digital impressions at the crown preparation margin using different scanning resolutions of a specific intraoral scanner system. An all-ceramic crown preparation of a mandibular first molar was constructed in a typodont, and a scan (n = 3) was created with an industrial-grade laboratory scanner (3Shape D2000) as the control. Digital impressions were obtained with an intraoral scanner (3Shape TRIOS 3) under three settings—high resolution (HR), standard resolution (SR), and combined resolution (SHR). Comparative 3D analysis of scans was performed with Geomagic Control X software to measure the discrepancy between intraoral scans and the control scan along the preparation finish line. The scan time and number of images captured per scan were recorded. Statistical analysis was performed by one-way ANOVA, two-way repeated measures ANOVA, Pearson’s correlation, and Dunnett’s T3 test (α = 0.05). Significant differences were observed for scan time and for number of images captured among scan resolution settings (α < 0.05). The scan time for the SR group was, on average, 34.2 s less than the SHR group and 46.5 s less than the HR group. For discrepancy on the finish line, no significant differences were observed among scanning resolutions (HR: 31.5 ± 5.5 μm, SHR: 33.2 ± 3.7 μm, SR: 33.6 ± 3.1 μm). Significant differences in discrepancy were observed among tooth surfaces, with the distal surface showing the highest discrepancies. In conclusion, the resolution of the intra-oral scanner is primarily defined by the system hardware and optimized for default scans. A software high-resolution mode that obtains more data over a longer time may not necessarily benefit the scan accuracy, while the tooth preparation and surface parameters do affect the accuracy.

Feedback mode preference and performance improvement with a handheld scanning device

This study was conducted to determine if combining feedback modes is beneficial for operator performance on a handheld scanning device within the services distribution industry. An industrial handheld scanner was used to examine four feedback conditions: 1) auditory; 2) auditory-visual; 3) auditory-tactile; 4) auditory-visual-tactile. Participants completed one trial under each experimental feedback condition during one experimental session. In each trial, participants scanned 50 boxes during a simulated box scanning task. Task completion time, completion time ranks, hit rate and false alarms were recorded and analysed. While the auditory-visual-tactile feedback combination produced the fastest task completion time, there was no significant improvement in operator performance between the four feedback settings tested. Through further understanding of the most preferred feedback modes, or combination of such, identification of improved scanner settings for this device and task can be made.

Feedback mode preference and performance improvement with a handheld scanning device

This study was conducted to determine if combining feedback modes is beneficial for operator performance on a handheld scanning device within the services distribution industry. An industrial handheld scanner was used to examine four feedback conditions: 1) auditory; 2) auditory-visual; 3) auditory-tactile; 4) auditory-visual-tactile. Participants completed one trial under each experimental feedback condition during one experimental session. In each trial, participants scanned 50 boxes during a simulated box scanning task. Task completion time, completion time ranks, hit rate and false alarms were recorded and analysed. While the auditory-visual-tactile feedback combination produced the fastest task completion time, there was no significant improvement in operator performance between the four feedback settings tested. Through further understanding of the most preferred feedback modes, or combination of such, identification of improved scanner settings for this device and task can be made.

Investigating the resolution capability of terrestrial laser scanners and its impact on the effective number of measurements

The resolution capability of terrestrial laser scanners does not necessarily equal the scanning resolution. Laser scanners can acquire the surrounding with a point distance of less than a millimeter on ten meters but the footprint size amounts to at least a few millimeters. Hence, neighboring laser spots overlap and they are correlated as they do not provide individual information about the object. This study provides a new methodology to determine the spatial expansion of the resolution capability of terrestrial laser scanners in the angular direction for distances up to 80 m. It is demonstrated that the Leica ScanStation P20 better resolves objects in the horizontal direction than in the vertical direction. Furthermore, the correlation of neighbored range measurements is empirically derived from the resolution capability and the effective number of measurements is determined. This provides a more realistic measure of the number of uncorrelated points in a scan. It is demonstrated that the resolution capability has a big impact on the correlation of the scan points as the number of uncorrelated points is reduced by up to 93% depending on the scanner settings.

Study of the effect of ultrasound scanner settings on the level of radiation intensity by the reciprocity method

In this paper, a new method of measuring the intensity on the radiating surface of the convex sensor of the ultrasonic diagnostic scanner EDAN U50 with its various settings was proposed, based on the reciprocity method using a hydrophone calibrated with a vibration simulator. The effect of the following scanner settings on the peak-spatial average time-averaged intensity in B-mode was experimentally investigated: ultrasonic radiation frequencies, number of focuses, and number of focus positions. It was found that changing the scanner settings significantly changes the temporal parameters of the emitted bursts of pulses, which affects the result of averaging intensity over time. At the same time, an increase in the number of focuses leads to a decrease in intensity, while an increase in the focus of the positions and frequency of the radiation, on the contrary, leads to its growth. Studies have shown that the intensity value exceeds the allowable values at almost any settings of the ultrasound scanner, with the exception of the ultrasound frequency of 2.5 MHz.

Application of Deep Learning Techniques for Characterization of 3D Radiological Datasets - A Pilot Study for Detection of Intravenous Contrast in Breast MRI.

Categorization of radiological images according to characteristics such as modality, scanner parameters, body part etc, is important for quality control, clinical efficiency and research. The metadata associated with images stored in the DICOM format reliably captures scanner settings such as tube current in CT or echo time (TE) in MRI. Other parameters such as image orientation, body part examined and presence of intravenous contrast, however, are not inherent to the scanner settings, and therefore require user input which is prone to human error. There is a general need for automated approaches that will appropriately categorize images, even with parameters that are not inherent to the scanner settings. These approaches should be able to process both planar 2D images and full 3D scans. In this work, we present a deep learning based approach for automatically detecting one such parameter: the presence or absence of intravenous contrast in 3D MRI scans. Contrast is manually injected by radiology staff during the imaging examination, and its presence cannot be automatically recorded in the DICOM header by the scanner. Our classifier is a convolutional neural network (CNN) based on the ResNet architecture. Our data consisted of 1000 breast MRI scans (500 scans with and 500 scans without intravenous contrast), used for training and testing a CNN on 80%/20% split, respectively. The labels for the scans were obtained from the series descriptions created by certified radiological technologists. Preliminary results of our classifier are very promising with an area under the ROC curve (AUC) of 0.98, sensitivity and specificity of 1.0 and 0.9 respectively (at the optimal ROC cut-off point), demonstrating potential usefulness in both clinical as well as research settings.

Image-plate sensitivity to x rays at 2 to 60 keV

The sensitivity of Fuji SR and MS image plates (IPs) used in x-ray spectrometers on OMEGA and the National Ignition Facility has been measured using two techniques. A set of radioisotopes has been used to constrain image-plate sensitivity between 6 and 60 keV, while a Manson source has been used to expose image plates to x rays at energies between 1.5 and 8 keV. These data have shown variation in sensitivity on the order of 5% for a given IP type and scanner settings. The radioisotope technique has also been used to assess IP fading properties for MS-type plates over long times. IP sensitivity as a function of scanner settings and pixel size has been systematically examined, showing variations of up to a factor of 2 depending on the IP type. Cross-calibration of IP scanners at different facilities is necessary to produce a consistent absolute sensitivity curve spanning the energy range of 2-60 keV.

C urve F usion

We introduce C urve F usion , the first approach for high quality scanning of thin structures at interactive rates using a handheld RGBD camera. Thin filament-like structures are mathematically just 1D curves embedded in R 3 , and integration-based reconstruction works best when depth sequences (from the thin structure parts) are fused using the object's (unknown) curve skeleton. Thus, using the complementary but noisy color and depth channels, C urve F usion first automatically identifies point samples on potential thin structures and groups them into bundles , each being a group of a fixed number of aligned consecutive frames. Then, the algorithm extracts per-bundle skeleton curves using L 1 axes, and aligns and iteratively merges the L 1 segments from all the bundles to form the final complete curve skeleton. Thus, unlike previous methods, reconstruction happens via integration along a data-dependent fusion primitive , i.e., the extracted curve skeleton. We extensively evaluate C urve F usion on a range of challenging examples, different scanner and calibration settings, and present high fidelity thin structure reconstructions previously just not possible from raw RGBD sequences.

Time domain model and experimental validation of non-contact surface wave scanner

This paper presents a time-domain model for the prediction of acoustic field in an air-coupled, non-contact, ultrasonic surface wave scanner, which includes an air-coupled Emitter, the Propagation space, and an air-coupled Receiver (EPR). The computation is divided into three steps, with each step being modeled in the time domain by its spatio-temporal transfer function. The latter are then used in turn, to find the pulse response of the overall system.The model takes the finite size of the aperture receiver, the attenuation in both air and the tested solid sample, as well as the electric response of the emitter-receiver set he into account. The attenuation is characterized by a causal time-domain Green’s function, allowing wideband attenuation of a lossy medium, obeying the power law αω=α0ωη,0⩽η⩽2, to be used. The model is implemented numerically using a Discrete Representation approach. It is then validated quantitatively by comparing the predicted acoustic field with experiment. The prediction error for three typical field features, the system’s impulse response, the on-axis field distribution, and the directivity pattern, is globally smaller than 3%.In order to obtain this high level of accuracy in the model, the parameters characterizing the solid sample used during the experiment were measured experimentally, with a specifically developed experimental setup.Overall, the proposed model is approximately 100 times faster than 3D FEM with an equivalent spatio-temporal resolution. In parallel, a simplified model is proposed, which neglects the attenuation in air and assumes the emitter inclination angle to be perfectly adjusted. This approach makes it possible to further shorten the computational time by a factor of about ten, whilst maintaining good accuracy. Thanks to its computational efficiency, the proposed model can be used to formulate various recommendations concerning the scanner settings, in particular the inclination angles of the emitter and receiver, and their distance from the sample.