Freehand 3D Research Articles

GPU-based volume reconstruction for freehand 3D ultrasound imaging.

Volume reconstruction plays an important role in improving image quality for freehand three-dimensional (3D) ultrasound imaging. The kernel regression provides an effective method for volume reconstruction in 3D ultrasound imaging, but it requires heavily computational time-cost. In this paper, a programmable graphic-processor-unit-(GPU) based fast kernel regression method is proposed for freehand 3D ultrasound volume reconstruction. The most significant aspect of our method is the adopting of powerful data-parallel computing capability of GPU to improve the overall efficiency. To produce higher image quality, the results of the kernel regression with various parameter settings is deeply investigated under the help of the fast implementation of the algorithm. Experimental results demonstrate that the computational performance of the proposed GPU-based method can be over 200 times faster than that on CPU. Better image quality for speckle reduction and details preservation can be obtained with the parameter setting of kernel window size of 5×5×5 and kernel bandwidth of 1.0.

3D intra-operative ultrasound and MR image guidance: pursuing an ultrasound-based management of brainshift to enhance neuronavigation.

Brainshift is still a major issue in neuronavigation. Incorporating intra-operative ultrasound (iUS) with advanced registration algorithms within the surgical workflow is regarded as a promising approach for a better understanding and management of brainshift. This work is intended to (1) provide three-dimensional (3D) ultrasound reconstructions specifically for brain imaging in order to detect brainshift observed intra-operatively, (2) evaluate a novel iterative intra-operative ultrasound-based deformation correction framework, and (3) validate the performance of the proposed image-registration-based deformation estimation in a clinical environment. Eight patients with brain tumors undergoing surgical resection are enrolled in this study. For each patient, a 3D freehand iUS system is employed in combination with an intra-operative navigation (iNav) system, and intra-operative ultrasound data are acquired at three timepoints during surgery. On this foundation, we present a novel resolution-preserving 3D ultrasound reconstruction, as well as a framework to detect brainshift through iterative registration of iUS images. To validate the system, the target registration error (TRE) is evaluated for each patient, and both rigid and elastic registration algorithms are analyzed. The mean TRE based on 3D-iUS improves significantly using the proposed brainshift compensation compared to neuronavigation (iNav) before (2.7 vs. 5.9mm; [Formula: see text]) and after dural opening (4.2 vs. 6.2mm, [Formula: see text]), but not after resection (6.7 vs. 7.5mm; [Formula: see text]). iUS depicts a significant ([Formula: see text]) dynamic spatial brainshift throughout the three timepoints. Accuracy of registration can be improved through rigid and elastic registrations by 29.2 and 33.3%, respectively, after dural opening, and by 5.2 and 0.4%, after resection. 3D-iUS systems can improve the detection of brainshift and significantly increase the accuracy of the navigation in a real scenario. 3D-iUS can thus be regarded as a robust, reliable, and feasible technology to enhance neuronavigation.

Integration of free-hand 3D ultrasound and mobile C-arm cone-beam CT: Feasibility and characterization for real-time guidance of needle insertion

This work presents development of an integrated ultrasound (US)–cone-beam CT (CBCT) system for image-guided needle interventions, combining a low-cost ultrasound system (Interson VC 7.5MHz, Pleasanton, CA) with a mobile C-arm for fluoroscopy and CBCT via use of a surgical tracker. Imaging performance of the ultrasound system was characterized in terms of depth-dependent contrast-to-noise ratio (CNR) and spatial resolution. US–CBCT system was evaluated in phantom studies simulating three needle-based procedures: drug delivery, tumor ablation, and lumbar puncture. Low-cost ultrasound provided flexibility but exhibited modest CNR and spatial resolution that is likely limited to fairly superficial applications within a ∼10cm depth of view. Needle tip localization demonstrated target registration error 2.1–3.0mm using fiducial-based registration.

Achilles tendon stress is more sensitive to subject-specific geometry than subject-specific material properties: A finite element analysis

This study used subject-specific measures of three-dimensional (3D) free Achilles tendon geometry in conjunction with a finite element method to investigate the effect of variation in subject-specific geometry and subject-specific material properties on tendon stress during submaximal isometric loading. Achilles tendons of eight participants (Aged 25–35years) were scanned with freehand 3D ultrasound at rest and during a 70% maximum voluntary isometric contraction. Ultrasound images were segmented, volume rendered and transformed into subject-specific 3D finite element meshes. The mean (±SD) lengths, volumes and cross-sectional areas of the tendons at rest were 62±13mm, 3617±984mm3 and 58±11mm2 respectively. The measured tendon strain at 70% MVIC was 5.9±1.3%. Subject-specific material properties were obtained using an optimisation approach that minimised the difference between measured and modelled longitudinal free tendon strain. Generic geometry was represented by the average mesh and generic material properties were taken from the literature. Local stresses were subsequently computed for combinations of subject-specific and generic geometry and material properties. For a given geometry, changing from generic to subject-specific material properties had little effect on the stress distribution in the tendon. In contrast, changing from generic to subject-specific geometry had a 26-fold greater effect on tendon stress distribution. Overall, these findings indicate that the stress distribution experienced by the living free Achilles tendon of a young and healthy population during voluntary loading are more sensitive to variation in tendon geometry than variation in tendon material properties.

Evaluation of Freehand B-Mode and Power-Mode 3D Ultrasound for Visualisation and Grading of Internal Carotid Artery Stenosis

BackgroundCurrently, colour-coded duplex sonography (2D-CDS) is clinical standard for detection and grading of internal carotid artery stenosis (ICAS). However, unlike angiographic imaging modalities, 2D-CDS assesses ICAS by its hemodynamic effects rather than luminal changes. Aim of this study was to evaluate freehand 3D ultrasound (3DUS) for direct visualisation and quantification of ICAS.MethodsThirty-seven patients with 43 ICAS were examined with 2D-CDS as reference standard and with freehand B-mode respectively power-mode 3DUS. Stenotic value of 3D reconstructed ICAS was calculated as distal diameter respectively distal cross-sectional area (CSA) reduction percentage and compared with 2D-CDS.ResultsThere was a trend but no significant difference in successful 3D reconstruction of ICAS between B-mode and power mode (examiner 1 {Ex1} 81% versus 93%, examiner 2 {Ex2} 84% versus 88%). Inter-rater agreement was best for power-mode 3DUS and assessment of stenotic value as distal CSA reduction percentage (intraclass correlation coefficient {ICC} 0.90) followed by power-mode 3DUS and distal diameter reduction percentage (ICC 0.81). Inter-rater agreement was poor for B-mode 3DUS (ICC, distal CSA reduction 0.36, distal diameter reduction 0.51). Intra-rater agreement for power-mode 3DUS was good for both measuring methods (ICC, distal CSA reduction 0.88 {Ex1} and 0.78 {Ex2}; ICC, distal diameter reduction 0.83 {Ex1} and 0.76 {Ex2}). In comparison to 2D-CDS inter-method agreement was good and clearly better for power-mode 3DUS (ICC, distal diameter reduction percentage: Ex1 0.85, Ex2 0.78; distal CSA reduction percentage: Ex1 0.63, Ex2 0.57) than for B-mode 3DUS (ICC, distal diameter reduction percentage: Ex1 0.40, Ex2 0.52; distal CSA reduction percentage: Ex1 0.15, Ex2 0.51).ConclusionsNon-invasive power-mode 3DUS is superior to B-mode 3DUS for imaging and quantification of ICAS. Thereby, further studies are warranted which should now compare power-mode 3DUS with the angiographic gold standard imaging modalities for quantification of ICAS, i.e. with CTA or CE-MRA.

The tendinopathic Achilles tendon does not remain iso-volumetric upon repeated loading: insights from 3D ultrasound.

Mid-portion Achilles tendinopathy (MAT) alters the normal three-dimensional (3D) morphology of the Achilles tendon (AT) at rest and under a single tensile load. However, how MAT changes the 3D morphology of the AT during repeated loading remains unclear. This study compared the AT longitudinal, transverse and volume strains during repeated loading of the tendinopathic AT with those of the contralateral tendon in people with unilateral MAT. Ten adults with unilateral MAT performed 10 successive 25 s submaximal (50%) voluntary isometric plantarflexion contractions with both legs. Freehand 3D ultrasound scans were recorded and used to measure whole AT, free AT and proximal AT longitudinal strains and free AT cross-sectional area (CSA) and volume strains. The free AT experienced higher longitudinal and CSA strain and reached steady state following a greater number of contractions (five contractions) in the tendinopathic AT compared with the contralateral tendon (three contractions). Further, free tendon CSA and volume strain were greater in the tendinopathic AT than in the contralateral tendon from the first contraction, whereas free AT longitudinal strain was not greater than that of the contralateral tendon until the fourth contraction. Volume loss from the tendon core therefore preceded the greater longitudinal strain in the tendinopathic AT. Overall, these findings suggest that the tendinopathic free AT experiences an exaggerated longitudinal and transverse strain response under repeated loading that is underpinned by an altered interaction between solid and fluid tendon matrix components. These alterations are indicative of accentuated poroelasticity and an altered local stress-strain environment within the tendinopathic free tendon matrix, which could affect tendon remodelling via mechanobiological pathways.

3D ultrasound imaging in frequency domain with 1D array transducer

Freehand three-dimensional (3D) ultrasound imaging usually uses a moving elevation focused one-dimensional (1D) transducer to construct a series of B-mode slice images, and then these B-mode slice images are combined to form a 3D volume image. This method has poor elevation resolution and low frame rate. In this paper, a novel 3D imaging method with a moving 1D array transducer is proposed. Different from the freehand 3D reconstruction, a linear array transducer without elevation focusing is used in the proposed method, and the 1D transducer transmits limited diffraction beams and receives echo signals repeatedly when moving along the elevation direction. After finishing the scanning, all the received signals are combined to construct the 3D image in Frequency domain with Fourier transform. Simulation results show that compared with the freehand imaging method, the new method can improve the elevation resolution significantly, and the resolution keeps almost fixed in the axial direction. The elevation resolution can be promoted to less than 2mm with the imaging depth 100mm by one transmission at each position. Besides, because only one transmission is needed at each position, the frame rate can be increased to some extent.

Knee Moment-Angle Characteristics and Semitendinosus Muscle Morphology in Children with Spastic Paresis Selected for Medial Hamstring Lengthening.

To increase knee range of motion and improve gait in children with spastic paresis (SP), the semitendinosus muscle (ST) amongst other hamstring muscles is frequently lengthened by surgery, but with variable success. Little is known about how the pre-surgical mechanical and morphological characteristics of ST muscle differ between children with SP and typically developing children (TD). The aims of this study were to assess (1) how knee moment-angle characteristics and ST morphology in children with SP selected for medial hamstring lengthening differ from TD children, as well as (2) how knee moment-angle characteristics and ST morphology are related. In nine SP and nine TD children, passive knee moment-angle characteristics and morphology of ST (i.e. fascicle length, muscle belly length, tendon length, physiological cross-sectional area, and volume) were assessed by hand-held dynamometry and freehand 3D ultrasound, respectively. At net knee flexion moments above 0.5 Nm, more flexed knee angles were found for SP compared to TD children. The measured knee angle range between 0 and 4 Nm was 30% smaller in children with SP. Muscle volume, physiological cross-sectional area, and fascicle length normalized to femur length were smaller in SP compared to TD children (62%, 48%, and 18%, respectively). Sixty percent of the variation in knee angles at 4 Nm net knee moment was explained by ST fascicle length. Altered knee moment-angle characteristics indicate an increased ST stiffness in SP children. Morphological observations indicate that in SP children planned for medial hamstring lengthening, the longitudinal and cross-sectional growth of ST muscle fibers is reduced. The reduced fascicle length can partly explain the increased ST stiffness and, hence, a more flexed knee joint in these SP children.

Freehand three-dimensional ultrasound imaging of carotid artery using motion tracking technology

Ultrasound imaging has been extensively used for determining the severity of carotid atherosclerotic stenosis. In particular, the morphological characterization of carotid plaques can be performed for risk stratification of patients. However, using 2D ultrasound imaging for detecting morphological changes in plaques has several limitations. Due to the scan was performed on a single longitudinal cross-section, the selected 2D image is difficult to represent the entire morphology and volume of plaque and vessel lumen. In addition, the precise positions of 2D ultrasound images highly depend on the radiologists’ experience, it makes the serial long-term exams of anti-atherosclerotic therapies are difficult to relocate the same corresponding planes by using 2D B-mode images. This has led to the recent development of three-dimensional (3D) ultrasound imaging, which offers improved visualization and quantification of complex morphologies of carotid plaques. In the present study, a freehand 3D ultrasound imaging technique based on optical motion tracking technology is proposed. Unlike other optical tracking systems, the marker is a small rigid body that is attached to the ultrasound probe and is tracked by eight high-performance digital cameras. The probe positions in 3D space coordinates are then calibrated at spatial and temporal resolutions of 10μm and 0.01s, respectively. The image segmentation procedure involves Otsu’s and the active contour model algorithms and accurately detects the contours of the carotid arteries. The proposed imaging technique was verified using normal artery and atherosclerotic stenosis phantoms. Human experiments involving freehand scanning of the carotid artery of a volunteer were also performed. The results indicated that compared with manual segmentation, the lowest percentage errors of the proposed segmentation procedure were 7.8% and 9.1% for the external and internal carotid arteries, respectively. Finally, the effect of handshaking was calibrated using the optical tracking system for reconstructing a 3D image.

3D freehand ultrasound reconstruction using a piecewise smooth Markov random field

In this paper, we introduce a novel three-dimensional (3D) reconstruction framework for ultrasound images using a piecewise smooth Markov random field (MRF) model from irregularly spaced B-scan images obtained by freehand scanning. Freehand 3D ultrasound imaging is a useful system for various clinical applications, including image-guided surgeries and interventions, as well as diagnoses, due to the variety of its scan ranges and relatively low cost. The reconstruction process performs a key role in this system because its sampling irregularities may cause undesired artifacts, and ultrasound images generally suffer from noise and distortions. However, traditional approaches are based on simple geometric interpolations, such as pixel-based or distance-weighted methods, which are sensitive to sampling density and speckle noise. These approaches generally have an additional limitation of smoothing objects boundaries. To reduce speckle noise and preserve boundaries, we devised a piecewise smooth (PS) MRF model and developed its optimization algorithm. In our framework, we can easily apply an individual noise level for each image pixel, which is specified by the characteristics of an ultrasound probe, and possibly, the lateral and axial positions of an image. As a result, the reconstructed volume has sharp object boundaries with reduced speckle noise and artifacts. Our PS-MRF model provides simple segmentation results within a reconstruction framework that is useful for various purposes, such as clear visualization. The corresponding optimization methods have also been developed, and we tested a virtual phantom and a physical phantom model. Experimental results show that our method outperforms existing methods in terms of interpolation and segmentation accuracy. With this method, all computations can be performed with practical time consumption and with an appropriate resolution, via parallel computing using graphic processing units.

The reliability and validity of a clinical 3D freehand ultrasound system

Background and ObjectiveAcquiring large anatomical volumes in a feasible manner is useful for clinical decision-making. A relatively new technique called 3D freehand ultrasonography is capable of this by combining a conventional 2D ultrasonography system. Currently, a thorough analysis of this technique is lacking, as the analyses are dependent on the software implementation details and the choice of measurement systems. Therefore this study starts by making this implementation available under the form of an open-source software library to perform 3D freehand ultrasonography. Following that, reliability and validity analyses of extracting volumes and lengths will be carried out using two independent motion-tracking systems. MethodsA PC-based ultrasonography device and two optical motion-tracking systems were used for data acquisition. An in-house software library called Py3DFreeHandUS was developed to reconstruct (off-line) the corresponding data into one 3D data set. Reliability and validity analyses of the entire experimental set-up were performed by estimating the volumes and lengths of ground truth objects. Ten water-filled balloons and six cross-wires were used. Repeat measurements were also performed by two experienced operators. ResultsThe software library Py3DFreeHandUS is available online, along with the relevant documentation. The reliability analyses showed high intra- and inter-operator intra-class correlation coefficient results for both volumes and lengths. The accuracy analysis revealed a discrepancy in all cases of around 3%, which corresponded to 3 ml and 1 mm for volume and length measurements, respectively. Similar results were found for both of the motion-tracking systems. ConclusionsThe undertaken analyses for estimating volume and lengths acquired with 3D freehand ultrasonography demonstrated reliable design measurements and suitable performance for applications that do not require sub-mm and -ml accuracy.

Reliability analysis on clinical relevant muscle morphology parameters extracted from 3D freehand Ultrasonography in healthy and pathological gastrocnemius

Reliability analysis on clinical relevant muscle morphology parameters extracted from 3D freehand Ultrasonography in healthy and pathological gastrocnemius

Validation of 2D ultrasound technique for tracking the medial gastrocnemius muscle-tendon junction by means of 3D Freehand Ultrasonography

Validation of 2D ultrasound technique for tracking the medial gastrocnemius muscle-tendon junction by means of 3D Freehand Ultrasonography

EP 74. Comparison of freehand B-mode and power-mode 3D ultrasound for visualisation and grading of internal carotid artery stenosis

Currently, colour-coded duplex sonography (2D-CDS) is clinical standard for detection and grading of ICAS (1, 2) as a relevant risk factor for ischaemic stroke (3, 4). Unlike angiographic imaging modalities, 2D-CDS assesses ICAS by its haemodynamic effects rather than luminal changes1. Therefore, aim of this study was to evaluate freehand 3D ultrasound (3DUS) for direct visualisation and quantification of ICAS. 37 patients with 43 ICAS were examined with 2D-CDS as reference standard and with freehand B-mode respectively power-mode 3DUS. For 3DUS Curefab CS (Curefab Technologies GmbH, Munich, Germany) was used. While 3DUS scanning was done by one examiner, ICAS were manually reconstructed within the virtual 3D-volume and graded by 2 physicians. Stenotic value of 3D reconstructed ICAS was assessed by calculating distal diameter and distal cross-sectional area (CSA) reduction percentage. There was a trend but no significant difference in successful 3D reconstruction of ICAS between B-mode and power mode (examiner 1 [Ex1] 81% versus 93%, examiner 2 [Ex2] 84% versus 88%). Interrater reliability (IR) was best for power-mode 3DUS and assessment of stenotic value as distal CSA reduction percentage (intraclass correlation coefficient [ICC] 0.90) followed by power-mode 3DUS and distal diameter reduction (ICC 0.81). IR was poor for B-mode 3DUS (ICC, distal CSA reduction 0.36; distal diameter reduction 0.51). In comparison to 2D-CDS intermethod reliability was clearly better for power-mode 3DUS (ICC, distal diameter reduction: Ex1 0.85, Ex2 0.78; ICC, distal CSA reduction: Ex1 0.63, Ex2 0.57) than for B-mode 3DUS (ICC, distal diameter reduction: Ex1 0.40, Ex2 0.52; ICC, distal CSA reduction: Ex1 0.15, Ex2 0.51). For power-mode 3DUS (distal diameter reduction) positive predictive value for differentiation between moderate and high-grade ICAS was 0.81 (Ex1) and 0.76 (Ex2) while negative predictive value was 0.92 (Ex1) and 0.91 (Ex2). Power-mode 3DUS is superior to B-mode 3DUS for imaging and quantification of ICAS. It might ideally complement 2D-CDS as initial vascular diagnostic in stroke patients and could be a simple alternative for more invasive and time-consuming imaging modalities like computed tomography angiography or contrast-enhanced magnetic resonance angiography.

Regional three-dimensional deformation of human Achilles tendon during conditioning.

Our understanding of invivo Achilles tendon (AT) conditioning is limited to two-dimensional ultrasound measures of longitudinal deformation of the whole tendon. This study investigated the regional three-dimensional (3D) deformation of the AT during conditioning. Eighteen ATs were scanned using 3D freehand ultrasound during 10 successive 25s submaximal (50%) voluntary isometric plantarflexion contractions. Longitudinal strain was assessed for the whole AT, aponeurosis, and free AT and transverse strain was assessed for the proximal-, mid-, and distal-portions of the free AT. Longitudinal conditioning of the whole AT was primarily driven by creep response of the free AT and transverse conditioning was greatest for the mid-portion of the free AT. Whole and free AT longitudinal strain increased up to the third contraction and were accompanied by a corresponding reduction in free AT cross-sectional area (CSA) strain in proximal-, mid-, and distal-portions. No significant changes in aponeurosis strain or tendon volume were detected between contractions. These findings suggest that conditioning alters free AT shape, with increased tendon length attained at the expense of reduction in free AT CSA. Although AT experiences different amounts of strain in different regions, the number of contractions required to reach steady-state strain during conditioning is uniform throughout the tendon.

Subject-specific and respiration-corrected 4D liver model from real-time ultrasound image sequences

Respiratory liver motion is one of the major issues affecting the liver interventions, such as biopsy and ablation, in the clinical practice. Traditional 4D liver model methods using magnetic resonance imaging (MRI) or computed tomography (CT) data sets for motion compensation to solve this problem have proved time-consuming and costly. In addition, the widely used freehand 3D ultrasound techniques, which lack respiratory corrections, cannot effectively track breathing-induced liver motion. On the other hand, clinical solutions are straightforward but present restrictions for patients. Motivated by both the technical and clinical needs, we propose a novel method for creating a sequence of subject-specific and respiration-corrected 3D ultrasound (US) images, from multiple robotic-assisted-swept 2D US image sequences. Moreover, we also observed the motion difference between subjects on the generated 4D model. The results of quantitative evaluation on the accuracy of the models show that the overlap ratio of the liver boundary between the generated model and ground truth at end of exhale and end of inhale phases were 0.90 and 0.89, respectively. The overall distance error of pinpointed landmarks was 2.44 mm, which is within the acceptance range of clinical applications. Therefore, we conclude that the reconstructed 3D image sequence can capture the moving liver during a half respiratory cycle, and the proposed method is feasible to visualise the 3D liver motion. The clinicians who worked with us in this study also suggest that this preoperative subject-specific motion information could help them to diagnose the existence, or determine the possible position, of a tumour.

Comparison of Freehand 3D and 4D Ultrasound to Measure Thrombus Volume in Patients with Acute Deep Vein Thrombosis

Objectives Current imaging techniques have limited ability to quantify thrombus burden in patients with acute deep vein thrombosis (DVT). There is limited ability to objectively measure clot progression or clot resolution over time and with treatment. These critical measures are important determinants of optimal treatment and therapeutic success for thrombolysis or thrombectomy treatment of DVT. In this study, we evaluated the feasibility and reliability of a method of quantifying thrombus volume using freehand three-dimensional (3D) ultrasonography and compared it to a commercially available four-dimensional (4D) transducer for acquiring and analyzing 3D ultrasound images to measure thrombus volume. Methods We studied 10 consecutive patients with a first episode of acute DVT to compare the performance of freehand 3D with 4D transducer imaging (group 1). In addition, we studied five patients with thrombus length >3.8 cm to evaluate the performance of freehand 3D imaging in larger clots (group 2). Scanning was performed with routine grayscale, color-flow, and power-Doppler imaging; followed by freehand 3D and 4D transducer volumetric imaging. Postprocessing was performed using the Stradwin software to measure thrombus volume after each scan. Results The age of patients was 54 ± 13 years (mean ± SD), 73.3% of DVTs were in the upper extremity, and 60% of the patients received anticoagulation. The mean thrombus volume in group 1 measured by freehand 3D and 4D probes was 0.49 ± 0.40 cm3 and 0.52 ± 0.46 cm3, respectively, and there was no significant difference between the measurements ( t = −1.57, p = 0.15). Regression analysis demonstrated good agreement between the two imaging modalities ( R2 = 0.99, regression slope = 1.13, p < 0.001). The freehand 3D technique was also successfully used to quantify thrombus volume (4.71 ± 3.28 cm3) in five patients with DVT >3.8 cm in length (group 2). Conclusions The described freehand 3D imaging protocol reliably measures thrombus volume in patients with acute DVT when compared with the commercially available 4D transducer. The freehand 3D method has the additional benefit of measuring thrombus volume in DVTs >3.8 cm in length. The method is convenient and can be used in clinical practice.

Sentinel Lymph Node Detection by 3D Freehand Single-Photon Emission Computed Tomography in Early Stage Breast Cancer.

We herein present our first experience obtained by 3D freehand single-photon emission computed tomography (SPECT) (F-SPECT) guidance for sentinel lymph node detection (SLND) in two patients with early stage breast cancer. F-SPECT guidance was carried out using one-day protocol in one case and by the two-day protocol in the other one. SLND was performed successfully in both patients. Histopathologic evaluation showed that the excised nodes were tumor negative. Thus, patients underwent breast-conserving surgery alone.

Real-time freehand 3D ultrasound imaging

Real-time freehand 3D ultrasound (US) has attracted much attention in clinical practices because it provides interactive feedback to help the clinicians acquire not only high-quality images but also timely information of the scanning area, which is necessary in intraoperative examinations. In this study, we developed a real-time freehand 3D US imaging system which can obtain volume reconstruction and visualisation during data acquisition at real-time level. Based on two popular algorithms, i.e. squared distance weighted interpolation (SDW) and Bezier interpolation, we designed corresponding parallel computing methods that were implemented on the graphics processing unit (GPU) to incrementally reconstruct and display the tissues being scanned using Visualisation toolkits (VTK). With a typical B-scan image size of 302 × 268 at an acquisition rate of 25 Hz and a preset volume size of 90 × 81 × 192, the system achieved an incremental reconstruction-visualisation rate of up to 32 frames/s and 119 frames/s for the SDW and Bezier algorithms, respectively, achieving the real-time 3D US.

Quantitative Assessment of Variational Surface Reconstruction from Sparse Point Clouds in Freehand 3D Ultrasound Imaging during Image-Guided Tumor Ablation

Surface reconstruction for freehand 3D ultrasound is used to provide 3D visualization of a VOI (volume of interest) during image-guided tumor ablation surgery. This is a challenge because the recorded 2D B-scans are not only sparse but also non-parallel. To solve this issue, we established a framework to reconstruct the surface of freehand 3D ultrasound imaging in 2011. The key technique for surface reconstruction in that framework is based on variational interpolation presented by Greg Turk for shape transformation and is named Variational Surface Reconstruction (VSR). The main goal of this paper is to evaluate the quality of surface reconstructions, especially when the input data are extremely sparse point clouds from freehand 3D ultrasound imaging, using four methods: Ball Pivoting, Power Crust, Poisson, and VSR. Four experiments are conducted, and quantitative metrics, such as the Hausdorff distance, are introduced for quantitative assessment. The experiment results show that the performance of the proposed VSR method is the best of the four methods at reconstructing surface from sparse data. The VSR method can produce a close approximation to the original surface from as few as two contours, whereas the other three methods fail to do so. The experiment results also illustrate that the reproducibility of the VSR method is the best of the four methods.