Ultrasound in Med. & Biol., Vol. 41, No. 1, pp. 309–316, 2015 Copyright Ó 2015 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/$ - see front matter

http://dx.doi.org/10.1016/j.ultrasmedbio.2014.08.023

d

Original Contribution HIGH-FREQUENCY ULTRASOUND AS AN OPTION FOR SCANNING OF PREPARED TEETH: AN IN VITRO STUDY FABRICE CHUEMBOU PEKAM,* JULIANA MAROTTI,y STEFAN WOLFART,y JOACHIM TINSCHERT,y KLAUS RADERMACHER,* and STEFAN HEGER*z * Chair of Medical Engineering, RWTH Aachen University, Aachen, Germany; y Department of Prosthodontics and Biomaterials, Medical Faculty, RWTH Aachen University, Aachen, Germany; and z Institute for Biomedical Engineering, Mannheim University of Applied Science, Manheim, Germany (Received 30 March 2014; revised 25 August 2014; in final form 26 August 2014)

Abstract—Because of its ability to non-invasively capture hard structures behind soft tissue, high-frequency ultrasound (HFUS)-assisted microscanning could be a patient-friendly and promising alternative for digitization of prepared teeth. However, intra-oral HFUS microscanners for taking digital impressions of prepared teeth are still not available in the clinical setting. Because working range, scanner size, scanning time, surface reconstruction accuracy and costs are major factors in such a system, our overall objective is to minimize hardware efforts and costs while maintaining the accuracy of the surface-reconstructed tooth model in the range 50 mm. In the work described here, we investigated the accuracy of tooth impression taking using a single-element HFUS microscanner with only three translational degrees of freedom under the restriction that only one occlusal scan is performed per tooth. As in favor of time and scanning efforts the data density is expected to be low, the surface reconstruction process is linked to a model-based surface reconstruction approach using a thin spline robust point matching algorithm to fill data gaps. A priori knowledge for the model is generated based on the original HFUS measurement data. Three artificial teeth and one human molar were prepared and scanned using an extra-oral HFUS laboratory microscanner that was built to test and evaluate different scanning setups. A scanner with three translational degrees of freedom was used to scan the teeth from an occlusal direction. After application of the proposed thin-spline robust point matching algorithm-based reconstruction approach, reconstruction accuracy was assessed by comparing the casts with a control group scanned with an extra-oral laser-scanning system. The mean difference between the reconstructed casts and the optical control group was in the range 14–53 mm. The standard deviation was between 21 and 52 mm. This let us assume that the suggested approach can help to decrease hardware efforts while maintaining the robustness of the 3-D surface reconstruction process for future HFUS-based intra-oral scanners. (E-mail: [email protected]) Ó 2015 World Federation for Ultrasound in Medicine & Biology. Key Words: Tooth impression taking, High-frequency ultrasound, Extra-oral 3-D ultrasound microscanning, Dental ultrasound, Dental prosthetics, Thin-spline robust point matching algorithm segmentation.

clear occlusal-facing ridge at the base of the crown. After preparation, one of the next steps in tooth restoration work flow is 3-D tooth impression taking. A precise and accurate impression, particularly at the preparation margin, is of crucial importance to the final fit of the crown, as well as to the health of the surrounding periodontal area (Wassell et al. 2002). Clinically, the fit of the crown is in the range 50–200 mm (Hunter and Hunter 1990; Mehl et al. 2009; Reich et al. 2005; Wolfart et al. 2003). However, conventional dental impression methods are still based on elastomer casting compounds followed by manufacture of gypsum models representing an analogue positive imprint of the prepared tooth as well as the gum situation. Nevertheless, these techniques are time consuming, potentially less reliable and user dependent. A limitation

INTRODUCTION Over the last decade, computer-aided design/computeraided manufacturing of highly durable and aesthetically satisfying restorations has become the center of attention. The first step in the restoration process is the preparation of the tooth by the dentist. The most frequently applied preparation type is the chamfer preparation. It enables a

Address correspondence to: Fabrice Chuembou Pekam, Helmholtz Institute for Biomedical Engineering, RWTH Aachen University, Pauwelsstrasse 20, 52074 Aachen, Germany. E-mail: chuembou@hia. rwth-aachen.de Conflicts of Interest: The authors have indicated that they have no conflicts of interest regarding the content of this article. 309

310

Ultrasound in Medicine and Biology

of this technique may be a less accurate fit of the final crown accompanied by an increased risk of periodontitis (Steele et al. 2002; Wassell et al. 2002) or caries. Furthermore, the detection of subgingival preparation margins demands retraction of the gingiva before impression taking, which can be achieved by retraction cords, electrosurgery or laser surgery. Such techniques usually cause the patient discomfort and bear a risk of permanent gingival retraction (Pathale et al. 2010). The benefits of an all-ceramic restoration (Luthardt et al. 2002) have, among other things, led to the development of computer-aided manufacturing (CAM) techniques for dental restorations. CAM, in turn, contributed to the establishment of systems for computer-aided design (CAD) (Palin and Burke 2005) and, particularly, the development of digital impression-taking systems such as intra-oral optical scanners (Babayoff 2004; Fisker et al. 2010; Rohaly 2006; Schmidt 2010). Although this technology is increasing in accuracy, digitization of subgingival preparation margins without uncovering the gingiva remains an unsolved problem, because optical waves can barely penetrate gingival tissue. Moreover, some of the optical systems need to be powdered using a biocompatible titanium dioxide powder to improve the reflectivity of the tooth surface (Palin and Burke 2005). This powder usually causes surface errors in the CAD model after digital reconstruction (Logozzo et al. 2011). As a non-invasive imaging modality, high-frequency ultrasound (HFUS)-based microscanning could be an alternative for intra-oral scanning of prepared teeth with high resolution. HFUS enables imaging of structures in the micrometer range and can penetrate soft tissue and fluids without causing physical or biological damage to the patient (Marotti et al. 2013b; Sun et al. 2007; Vogt and Ermet 2008). A commercial diagnostic ultrasound device for intra-oral scanning is, to our knowledge, not available (Marotti et al. 2013b). Because the accuracy of fit of the dental prosthesis depends strongly on the accuracy of the impression taking, tooth impression taking based on ultrasound requires a system with high spatial resolution. The full pulse width at half-maximum (FWHM) describes the ability of the system to distinguish between two points and is the standard measure used to describe this resolution. In the case of diffraction-limited systems and for a spherically focused transducer, the FWHMs in the lateral and axial directions are given by dlat 5 lm 3f#

(1)

dax 5 ð2c3ln 2Þ=pB

(2)

lm is the average wavelength, which corresponds to the sound velocity c of the medium divided by the center

Volume 41, Number 1, 2015

frequency fm of the traveling pulse in the transmit/receive path; f# is the focal ratio; and B is the frequency bandwidth of the system’s transfer function limited mainly by the ultrasound probe. If it is assumed that f# 5 2, fm should be at least 50 MHz to achieve a lateral FWHM resolution in the range of 60 mm. This yields an axial FWHM resolution of approximately 18 mm in the case of B 5 0.75fm. If frequency-dependent attenuation of coupling medium and gingiva is taken into account, the center frequency should be even higher. In the lower frequency range (,20 MHz), multichannel array systems are successfully established in diagnostic ultrasound, enabling electronic scanning, focusing and steering of transmitted and received waves. However, the development of their high-frequency counterpart is still a challenging and expensive task. Within the framework of the intra-oral tooth impression based on ultrasound project, a novel HFUS microscanner for intra-oral dental impression taking is being developed (Vollborn et al. 2012a, 2012b). The scanner should be able to scan prepared teeth within the context of a tooth restoration process. For miniaturization purposes as well as cost reduction, the suggested scanning concept is based on a 2-degree of freedom (DOF) serial kinematic and an integrated focused single-element HFUS transducer with at least 50 MHz. During the scanning procedure, the ultrasound sensor is moved laterally over the image volume, for example, parallel to the occlusal surface of the tooth with the beam axis aligned with respect to the occlusal tooth’s surface normal. However, the imaging capability of a focused single-element transducer suffers from a limited depth of field. Hence, both the lateral resolution and the signal-to-noise ratio drop far from the focus. A solution to achieve almost homogeneous resolution over all depths of interest can be given by the synthetic aperture focusing technique (SAFT) (Chuembou Pekam et al. 2012; Frazier and O’Brien 1998; Passmann and Ermert 1996). Preliminary promising investigations of this approach have been made, and the development of a time- and memory-efficient 3-D SAFT algorithm is part of our ongoing work. An alternative to overcome this problem is to add one more degree of freedom to adjust mechanically the focus area, so that the system builds a 3-DOF scanner unit. Another challenge of this scanning concept is that reflections from the steep side walls of prepared teeth are usually very weak compared with the strong echoes from the occlusal surface. This is due to the glossy surface of the prepared tooth if compared with the wavelength of HFUS. As a consequence, noisy or missing data make the segmentation and meshing process challenging and vulnerable to error. It might therefore be challenging to adequately reconstruct the tooth’s surface.

High-frequency US for scanning prepared teeth d F. CHUEMBOU PEKAM et al.

In a previous project, a laboratory 5-DOF extra-oral ultrasound scanner offering overall scanning accuracy in the range of 5 mm was developed. This system was built mainly to test different scan strategies as well as algorithms for tooth surface reconstruction on the basis of HFUS data. Within a study, 3 1 1-DOF (1-DOF manually) has been used to generate four partial scans of teeth, with each scan made from a different direction (manual DOF). Each 3-DOF scan covered a 2-DOF area scan plus additional variation of scanning depth, yielding a 3-D scan volume. The four partial 3-DOF scans were matched together, and the entire data model was segmented using a threshold-based segmentation approach. The final point clouds yielded a homogeneous spread of surface points over the entire model such that standard meshing tools could have been applied (Heger et al. 2013; Marotti et al. 2013a). Regardless of the high accuracy reported in the study, in an intra-oral environment, either repositioning of the scanner between partial scans or development of a scanner with a much more complex scanner kinematic is required. Both solutions will lead to higher scanning efforts, longer scanning times and higher costs and, hence, are out of the scope of this study. Our goal in this study was to develop and evaluate an algorithm that is able to close larger data gaps under noisy conditions. This allows us to develop and improve a scanning system offering a less complex kinematic, similar to the 2-DOF intra-oral scanner described by

311

Vollborn et al. (2012b). As part of this study we investigated the accuracy of 3-DOF HFUS extra-oral microscanning of prepared teeth in combination with the thin-plate spline robust point matching (TPS-RPM) segmentation method (Chui and Rangarajan 2003) when only one occlusal scanning direction is used for tooth digitization. The study was based on the same extraoral scanning system described previously by Heger et al. (2013). This has the advantage that the results of this study are comparable to the results obtained by 4DOF scanning based on different scanning directions. Unlike the introduced 2-DOF scanning concept, the use of SAFT to extend focus depth is replaced within this study by adding a third DOF. METHODS Data acquisition Three artificial teeth (KaVo Dental, Biberach an der Riss, Germany) and one human molar (K1, K2, K3 and H, respectively) were prepared by a dentist and represent a real clinical case. The preparation type was the chamfer preparation. The human tooth was obtained from a patient who consented to its use for research and educational purposes. The four teeth were scanned in a water tank with a 3-DOF (33 translation) serial kinematic laboratory extra-oral microscanner (Fig. 1) and an integrated

Fig. 1. Ultrasound scanning setup (top) and tooth samples (bottom).

312

Ultrasound in Medicine and Biology

Fig. 2. Steps in the segmentation process. STL 5 Standard triangulation language; TPS-RPM 5 thin spline robust point matching algorithm; US 5 ultrasound.

single-element focused transducer (75-MHz center frequency, 78% frequency bandwidth, 6.35-mm aperture diameter, F# 5 2). The theoretical and measured FWHM lateral resolutions of the ultrasound probe were about 40 and 60 mm at the focus depth. This difference is due to frequency-dependent attenuation, which leads to a downshift in the center frequency. The transducer was connected to a commercial highfrequency pulser–receiver (300-MHz bandwidth). The tooth samples were positioned such that the beam axis was aligned with respect to the occlusal tooth’s surface normal. Figure 2 illustrates the setup of the tooth digitization process. Preliminary investigations indicated that the frequency range of the transducer used allows the detection of subgingival tooth preparation margins for gingiva up to 1 mm thick. Vandana and Savitha (2005) reported that the average human gingiva thickness is between 0.93 and 1.78 mm. In ongoing work, the maximal penetrable

Volume 41, Number 1, 2015

gingiva thickness using our system is being investigated. However, the digitization of subgingivally prepared teeth was out of the scope of this study and will be the subject of future work. To achieve the best possible scanning resolution, volumetric data on the tooth samples were acquired by moving the single-element transducer in azimuth and in the elevation direction with a step size of 20 mm and in the axial direction with a step size of 200 mm. For each transducer position, a pulse-echo ultrasound measurement was carried out. At the boundary between water and tooth, the signal is reflected according to the reflection law. Part of the backscattered echo is captured by the same transducer and recorded via the ultrasound system. After being amplified and prefiltered using a bandpass filter (30–150 MHz), the recorded ultrasound data are digitized with an analogue-to-digital converter (400 MS/s, 12 bit) and then stored on a computer. Tooth surface reconstruction After envelope detection of the recorded radiofrequency data using a Hilbert filter, a threshold filtering was applied on the 3-D A-mode data set to suppress noise. Then a 3-D point cloud data was generated by calculating the coordinate of the maximum in each Amode data line. Here it was assumed that the maximum reflection represents a tooth surface point. This assumption is based on the fact that the reflectivity of enamel in water in the case of perpendicular sound incidence is about 70%.

Fig. 3. Three-dimensional point cloud of the ultrasound scan of the four tooth samples before applying the thin spline robust point matching algorithm-based segmentation approach.

High-frequency US for scanning prepared teeth d F. CHUEMBOU PEKAM et al.

313

Fig. 4. Three-dimensional point cloud of the ultrasound scan of the four tooth samples after applying the thin-spline robust point matching algorithm-based segmentation approach.

The weak reflections of certain tooth surface points, particularly of the lateral surface of the tooth side walls, are a major challenge for the surface reconstruction process. An automated reconstruction algorithm must not only be accurate, but also be able to cope with noise and measurement errors as well as data gaps caused by weak or even missing echoes from the tooth’s side walls. Standard approaches like pixel-based and edge-based segmentation methods are not robust enough against the mentioned problems. We investigated the use of TPSRPM for feature-based registration of tooth surfaces (Chui and Rangarajan 2003). TPS-RPM is a non-rigid approach to find the optimal mapping between two sets of features. In comparison with rigid methods, TPSRPM can better deal with gaps or outliers as is the case here. To reconstruct the entire tooth surface, a 2-D segmentation approach based on the TPS-RPM algorithm was implemented. Our implementation of the TPSRPM algorithm requires an initial deformable model which was estimated based on the threshold segmented A-mode data lines. At first, a linear interpolation was applied in every 2-D slice of the 3-D point cloud to increase density and homogeneity of the initial point cloud. In the case of large information gaps (e.g., .1 mm) on the lateral tooth surface, intermediate points were also added by linear interpolation. Then, for each linear interpolated data slice, a corresponding segmentation model was generated by approximating the interpolated tooth contour using a polynomial fit. The so-generated initial segmentation model was used together with the threshold

segmented A-mode data set to automatically reconstruct the tooth surface with the TPS-RPM approach. By use of elastic transformation, the best fit of the model with the original data set was found under the condition of lowest energy consumption. Figure 2 describes the steps of the entire segmentation process. The accuracy of the reconstructed casts was accessed by calculating the discrepancy between the HFUS reconstructed casts and an optical control group acquired using a commercially available extra-oral optical scanner (D250, 3 Shape, Denmark). Extra-oral scanning systems are designed primarily for digitization of plaster models placed inside the scanner and hence make use of different angles of perspectives (e.g., eight or more) to capture the model entirely with the laser light. Later, the gathered data are matched together and meshed by the system automatically. Standard triangulation language (STL) data of the surface mesh are forwarded via a research user interface. The error distribution was given in terms of maximum, average and standard deviation. RESULTS Figure 3 illustrates the preparation margins and the occlusal surface of all tooth samples, accurately captured by the HFUS scan. As expected, the digitization of the lateral tooth surfaces was less successful (Fig. 3). In particular, the reflections coming from the lateral surfaces of the extracted human tooth were weaker, probably because the echogenicity of human tooth surfaces is inferior to that of model teeth. Moreover, Figure 3 revealed

314

Ultrasound in Medicine and Biology

Volume 41, Number 1, 2015

Fig. 5. Color difference maps of the deviation between the four thin spline robust point matching algorithm-based segmentations of ultrasound data sets and the corresponding optical scans of the teeth.

some outliers, which appeared despite the threshold filtering of the volumetric ultrasound data. The 2-D TPS-RPM-based segmentation approach was able to deal with the information gaps and the outliers. Reconstruction of the entire tooth surface was possible for all tooth samples (Fig. 4). In Figure 5, a color difference map illustrates the discrepancy in the digital imprints of the four samples acquired by ultrasound against the optical control group. The maximum deviation was 72 mm for K1, 109 mm for K2, 124 mm for K3 and 162 mm for the human tooth. The maximal average discrepancy (53 mm) and the maximal standard deviation (52 mm) were observed for the human cast. The average discrepancy and the standard deviation for the three artificial teeth were in the ranges 14–26 and 21–22 mm, respectively (Table 1). DISCUSSION In this study, a novel dental microscanning concept based on HFUS has been introduced as an alternative to

optically based digital impression taking of prepared teeth. Ultrasound has the ability to penetrate soft tissue and to capture structures such as subgingival preparation margins without invasive retraction of the gingiva. Thus, the proposed concept should be much less invasive than optical intra-oral scanning systems, which still require the retraction of subgingival margins. Our proposed mechanical scanning concept, however, is limited to one Table 1. Differences between the four ultrasound-based reconstructed tooth samples and the corresponding optical scans Mean difference (mm)

KaVo* model teeth K1 K2 K3 Human molar

Positive

Negative

Standard deviation

14 14 15 30

26 20 22 53

21 22 21 52

* KaVo Dental, Biberach an der Riss, German.

High-frequency US for scanning prepared teeth d F. CHUEMBOU PEKAM et al.

scanning direction, making it difficult to digitize the entire tooth surface. When scanning from an occlusal direction, as is the case here, steeply inclined specular reflectors, such as the tooth’s lateral surfaces, suffer from weak echogenicity. Hence, possible echoes are covered by noise and cannot be recovered. To overcome this problem and close data gaps caused by missing information, a TPS-RPM-based tooth surface reconstruction algorithm has been implemented and tested on ultrasound data sets of four teeth. On the basis of the proposed approach, the algorithm was able to automatically close all data gaps and to reconstruct entirely the digitized surfaces of all tooth samples. To assess the accuracy of the reconstructed casts, we compared the tooth surface reconstructions based on HFUS with extra-oral optical laser scans. In comparing the results of the model teeth with those for the human molar, the reconstruction accuracy of the latter was inferior. This correlates to the observation that the echogenicity of the than that of the model teeth. Because the uncertainty increases with the amount of information that must be estimated by the TPS-RPM-based segmentation approach, the final accuracy decreases. In a previous study, we evaluated the accuracy of ultrasound-based tooth surface reconstructions based on registered ultrasound data sets acquired from different scanning directions (Heger et al. 2013). Use of more than one scanning direction allows reduction of missing information and, hence, decreases reconstruction errors. Although this technique has been proven to have higher scanning accuracy, the scanning time and hardware expenses are higher as well (Heger et al. 2013). At this point it should be stated that unlike in the previous study, we did not make use of second harmonic imaging, which might improve the accuracy of our results further. In addition, it should be mentioned that the proposed algorithm requires a minimum amount of information close to the edge of the preparation margin. Although the teeth were prepared by a dentist under a realistic scenario and although edge detection was not a problem within this study, there might exist preparations where the edge of the margin is not visible in the occlusal scan or the data are not dense enough. In those cases, a 4-DOF scanning approach would be preferred. The reconstruction of the tooth surface was based on a 2-D segmentation approach. Hence, not all available information was taken into account. A 3-D segmentation approach could increase the accuracy of the entire reconstruction. For this study, the ultrasound scans were performed in a water tank. Water was used as a coupling medium for the ultrasound signal. This setup may not ideally reflect the clinical situation in the mouth, but it avoids additional sources of error. Furthermore, the self-developed extra-

315

oral HFUS system used for this study is not appropriate for intra-oral use. However, a clinical study and the miniaturization of the HFUS microscanner were out of the scope of this study and will be subject of future work. CONCLUSIONS This in vitro study illustrated that the accuracy of single-element transducer-based HFUS extra-oral microscanning of prepared teeth using a 3-DOF serial kinematic is not much worse than that of extra-oral 4-DOF scanning based on four different scanning directions followed by matching of the partial scans (Heger et al. 2013). Also, it is not much worse than that of a commercial extra-oral optical system, which is still viewed as the gold standard in tooth scanning. Comparison of the reconstructed casts of the three artificial teeth using the TPS-RPM image segmentation algorithm revealed an average discrepancy ,30 mm between the ultrasound scans and the optical reference group. The standard deviation was ,25 mm, which is close to clinical requirements. These results let us assume that scanning of prepared teeth using the proposed 2-DOF concept can be an alternative to optical scanning systems in the near future, especially in cases where preparation margins are covered by gingiva. Further, the proposed segmentation and reconstruction algorithm makes HFUS tooth microscanning more robust against noise and helps to close large data gaps. The HFUS microscanning approach described is applicable only to prepared teeth. Future work will focus on miniaturization and evaluation of the entire system. A miniaturized prototype of the proposed system using a smaller HFUS probe is described in Vollborn et al. (2012b) and will be evaluated soon. Investigation of the accuracy of subgingival scans and integration of the SAFT technique to reduce the number of degrees of freedom from 3 to 2 without loss of information and deterioration of the resolution are also part of our ongoing work. Acknowledgments—The contents of this article have been developed within a project granted by the German Federal Ministry of Education and Research (BMBF) under Grants 01 EZ0926 and 1 EZ0929.

REFERENCES Babayoff N. Imaging a three-dimensional structure by confocal focussing an array of light beams. Patent US 006697164 B1. 2004. Available at: http://www.google.com/patents/US6697164. Chuembou Pekam F, Vollborn T, Tinschert J, Wolfart S, Radermacher K, Heger S. Synthetic analysis of synthetic aperture focusing based on the virtual source point for high-frequency ultrasound imaging. In: Proceedings, 2012 IEEE International Ultrasonics Symposium. New York: IEEE; 2012. p. 1–4. Chui H, Rangarajan A. A new point matching algorithm for non-rigid registration. Comput Vis Image Understand 2003;89:114–141. € Fisker R, Ojelund H, Kjær R, Van Der Poel M, Qazi AA, Hollenbeck KJ. Focus scanning apparatus. Patent WO 002010145669 A1. Available at: http://ispub.com/IJMT/5/1/10082; 2010.

316

Ultrasound in Medicine and Biology

Frazier G, O’Brien W. Synthetic aperture technique with a virtual source element. IEEE Trans Ultrason Ferroelect Freq Contr 1998;45:196–207. Heger S, Vollborn T, Tinschert J, Wolfart S, Radermacher K. Accuracy assessment of high frequency 3D ultrasound for digital impressiontaking of prepared teeth. SPIE Proc Conf 2013;8675. Hunter AJ, Hunter AR. Gingival crown margin configurations A review and discussion: Part I. Terminology and widths. J Prosthet Dent 1990;64:548–552. Logozzo S, Franceschini G, Kilpel€a A, Caponi M, Governi L, Blois L. A comparative analysis of intraoral 3-D digital scanners for restorative dentistry. Internet J Med Technol 2011;5:1–22. Luthardt R, Weber A, Rudolph H, Schone C, Quaas S, Walter M. Design and production of dental prosthetic restorations: Basic research on dental CAD/CAM technology. Int J Comp Dent 2002;5:165–176. Marotti J, Heger S, Tinschert J, Kirsten A, Wolfart S. Ultrasound as a new technology for dental-scanning. In: Proceedings, International Association for Dental Research meeting, Seattle, Washington, USA, March 20–23, 2013a. Marotti J, Heger S, Tinschert J, Tortamano P, Chuembou F, Radermacher K, Wolfart S. Recent advances of ultrasound imaging in dentistry: A review of the literature. Oral Surg Oral Med Oral Pathol Oral Radiol 2013b;115:819–832. Mehl A, Ender A, M€ ormann W, Attin T. Accuracy testing of a new intraoral 3-D camera. Int J Comp Dent 2009;12:11–28. Palin W, Burke F. Trends in indirect dentistry: CAD/CAM technology. Dent Update 2005;32:566–572. Passmann C, Ermert H. A 100 MHz ultrasound imaging system for dermatologic and ophthalmologic diagnostics. IEEE Trans Ultrason Ferroelectr Freq Control 1996;43:545–552. Pathale S, Marawar P, Byakod B, Lagdive S, Kalburge J. Effect of retraction materials on gingival health: A histopathological study. J Indian Soc Periodontol 2010;14:35–39. Reich S, Wichmann M, Nkenke E, Proeschel P. Clinical fit of allceramic three-unit fixed partial dentures, generated with three different CAD/CAM systems. Eur J Oral Sci 2005;113:174–179.

Volume 41, Number 1, 2015 Rohaly J. Three-channel camera systems with non-collinear apertures. Patent US 7372642 B2. 2006. Available at: http://www.google. com/patents/US7372642. Schmidt V. 3D dental camera for recording surface structures of a measuring object by means of triangulation. Patent US 20110242281 A1. 2010. Available at: http://www.google.com/patents/US20110242281. Steele J, Nohl F, Wassell R. Crowns and other extra-coronal restorations: Occlusal considerations and articulator selection. Br Dent J 2002; 192:377–387. Sun L, Richard W, Cannata J, Feng C, Johnson J, Yen J, Shung K. A high-frame rate high-frequency ultrasonic system for cardiac imaging in mice. IEEE Trans Ultrason Ferroelectr Freq Control 2007; 54:1648–1655. Vandana KL, Savitha B. Thickness of gingiva in association with age, gender and dental arch location. J. Clin Periodontol 2005;32: 828–830. Vogt M, Ermet H. Limited-angle spatial compound imaging of skin with high-frequency ultrasound (20 MHz). IEEE Trans Ultrason Ferroelectr Freq Control 2008;55:1975–1983. Vollborn T, Chuembou F, Habor D, Tinschert J, Wolfart S, Radermacher K, Heger S. Concept and design for an intraoral ultrasonic microscanner. In: Proceedings, Automed—Automatisierungstechnische Verfahren f€ur die Medizin. Dusseldorf: VDI Verlag; 2012a. p. 39–40. Vollborn T, Habor D, Junk S, Radermacher K, Heger S. A voice-coil actuated ultrasound micro scanner for intraoral high resolution impression taking. IEEE Int Conf Intell Robots Syst 2012b;3571–3576. Wassell R, Barker D, Walls A. Crowns and other extra-coronal restorations: Impression materials and technique. Br Dent J 2002;192: 679–690. Wolfart S, Wegner SM, Al-Halabi A, Kern M. Clinical evaluation of marginal fit of a new experimental all-ceramic system before and after cementation. Int J Prosthodont 2003;16:587–592.

High-frequency ultrasound as an option for scanning of prepared teeth: an in vitro study.

Because of its ability to non-invasively capture hard structures behind soft tissue, high-frequency ultrasound (HFUS)-assisted microscanning could be ...
3MB Sizes 0 Downloads 9 Views