• Users Online: 285
  • Print this page
  • Email this page

 Table of Contents  
Year : 2020  |  Volume : 10  |  Issue : 4  |  Page : 239-248

Automatic evaluation of crown preparation using image processing techniques: A substitute to faculty scoring in dental education

1 Department of Dental Public Health, Dental Research Center, Dental Research Institute, Isfahan University of Medical Sciences, Khorramabad, Iran
2 Department of Computer Engineering, Faculty of Engineering, Lorestan University, Khorramabad, Iran
3 School of Electrical and Computer Engineering, Faculty of Engineering, University of Tehran, Tehran, Iran
4 Department of Prosthodontics, Dental Research Center, Dental Research Institute, Isfahan University of Medical Science, Isfahan, Iran
5 Department of Electrical Engineering, Sepahan Science and Technology Institute of Higher Education, Khorramabad, Iran
6 Medical Image and Signal Processing Research Center, School of Advanced Technologies in Medicine, Isfahan University of Medical Sciences, Khorramabad, Iran

Date of Submission17-Jan-2020
Date of Decision19-Feb-2020
Date of Acceptance11-Mar-2020
Date of Web Publication11-Nov-2020

Correspondence Address:
Dr. Rahele Kafieh
Medical Image and Signal Processing Research Center, School of Advanced Technologies in Medicine, Isfahan University of Medical Sciences, Isfahan
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/jmss.JMSS_5_20

Rights and Permissions

Background: This study presents a new and innovative experimental method, including software and its prerequisite instruments, to use image processing techniques for crown preparation analysis. Method: A platform was designed and constructed to take images from artificial teeth in different angles and directions and to process and analyze them by the proposed method to evaluate the quality and quantity of crown preparation. For each tooth, two series of images were taken from the artificial teeth before and after preparation, and image series were registered by two semi-automated and automated methods to transform them into one coordinate system. Region of interest was segmented by user interaction, and tooth region was segmented by substeps such as transformation to hue, saturation, and value color space, edge detection, morphology operations, and contour extraction. Finally, the amount and angle of crown preparation were computed and compared with standard measures to evaluate the quality of crown preparation. The proposed method was applied to a local dataset collected from Isfahan University of Medical Sciences. Results: Difference between the angle of crown preparation computed by the proposed method and that of the experts showed a mean absolute error of 7.17°. The correlation between the segmented regions by the proposed method and those of the experts was also evaluated by the Intersection over Union (IOU) criterion. The best and worst performances achieved in cases by IOU were 0.94 and 0.76, respectively. Finally, the segmentation results of the proposed method indicated an average IOU of 0.89 in all images. Conclusion: Students can use this method as an assessment tool in preclinical tooth preparation to compare their crown work with standard parameters.

Keywords: Assessment, computer-assisted image processing, dental education, grading, preclinical simulation, prosthodontic tooth preparation

How to cite this article:
Tahani B, Rashno A, Haghighi H, Monirifard R, Khomami HN, Kafieh R. Automatic evaluation of crown preparation using image processing techniques: A substitute to faculty scoring in dental education. J Med Signals Sens 2020;10:239-48

How to cite this URL:
Tahani B, Rashno A, Haghighi H, Monirifard R, Khomami HN, Kafieh R. Automatic evaluation of crown preparation using image processing techniques: A substitute to faculty scoring in dental education. J Med Signals Sens [serial online] 2020 [cited 2021 Jan 19];10:239-48. Available from: https://www.jmssjournal.net/text.asp?2020/10/4/239/300503

  Introduction Top

Teeth do not possess the regenerative ability found in most other tissues. Therefore, once enamel or dentin is lost, restorative materials must be used to reestablish form and function.[1] Crown preparation is essential before the restoration stage, and preparations must be based on fundamental principles to establish basic criteria to predict the success of care in the future.[2]

In dentistry, crown is one of the most widely used fixed restorations. The restoration process consists of three groups: full metal, metal-ceramic, and full ceramic crowns. A full metal crown is a full-coverage metal restoration with a low level of esthetic options. A metal-ceramic crown is a full-coverage cast metal crown covered with a layer of fused porcelain to reproduce the aspect of natural tooth in both shape and color.[3] All-ceramic crowns are the most pleasing esthetic restorations which mimic the original tooth color better than other restorative options.[4]

Such a restoration requires specific tooth reduction to provide space for metal and/or ceramic layers of crown along with tooth structure preservation. Crown preparation requires considerable tooth reduction, leads to appropriate shape and thickness, and guarantees natural tooth duplication and structural durability. Insufficient tooth reduction may lead to a compromised appearance of restoration, and overcontoured restorations may appear at expense of proper axial contour, which leads to periodontal diseases. Sufficient tooth removal is needed to provide a distinct fishing line to host the material.[5] The recommended reduction in literature for different materials and views of tooth ranges from 0.7 in full metal to 2 mm in all-ceramic crowns.[6]

An appropriate degree of taper for crown preparation is also another important aspect in crown preparation. A too-small taper may lead to unwanted undercuts, causing unfitted restoration and possible tooth decay or tooth loss. A too-large taper leads to frequent decementation of the crown. Textbooks in fixed prosthodontics often recommend a convergence angle (CA) of approximately 5° (4°–6°) as ideal and a range of 4°–14° as acceptable.[7],[8] However, these guidelines are difficult to follow clinically, and divergence from parallel might have to be as much as 12° to be observed clinically as a diverging surface. Mack observed that a minimal taper of 12° was necessary to ensure the absence of undercuts.[9] Goodacre proposed that the total CA should range from 10° to 20°.[10] Other researchers have recommended 10° and 16° CA based on laboratory studies.[8] The recommended convergence between opposing walls is 6°, which is shown to optimize the retention of the crown. One of the most important components of dental education is preclinical fixed prosthodontics. Achieving proper skills in this domain is vital for becoming the next dentist. Progress for students in this field depends on hard practice, and one of the effective ways to provide students with efficient practice is developing a self-examination system.

There are several methods to evaluate the quality and quantity of preparation. The most accurate method to evaluate crown preparation includes making a proper impression, producing stone die, carrying out a two-dimensional scan of it, and measuring the preparation thickness and CA.[2] This method is not practical in a preclinical laboratory for several students due to time restrictions and need for additional steps. The current feasible method to evaluate preparation is undercut detection by an expert using visual techniques. There is no doubt that subjective evaluation suffers from intersubjective differences and is a time-consuming process to be used, which suffers from the level of expertise and the grading methods used.[11]

In recent decades, as a result of continual developments in digital dental technology and to improve the traditional techniques of evaluation, dental education has started to use digital technologies in various preclinical trainings[12] including tooth preparation. However, many of the available techniques require complicated equipment and enough time to work.

Computer-aided systems that offer many feedback options and self-assessment ability are used in different dental schools around the world. The Simodont Dental Trainer (ACTA, Amsterdam, The Netherlands) enables students to develop their manual skills in a realistic virtual world while getting feedback on their decisions and abilities. Sirona (Salzburg, Austria) describes its prepCheck as an easy-to-learn application that is based on objective measurement procedures with the ability to document and analyze students' results.[13] However, studies conducted to evaluate its effectiveness in dental education revealed that some students complained about working with this system and believed that introducing prepCheck requires enough equipment and preparation time.[14],[15]

Recently, E4D Compare software program (Richardson, TX, USA) has also been introduced with the ability to evaluate the prepared teeth by the three-dimensional (3D) technique.[16] Although the instruments described above have been shown to have successful outcomes, they are based on high-tech and high-price scanner systems. Further, in these systems, the teeth prepared by students need to be compared with the ones prepared by their supervisors, which poses the evaluation at risk of human judgment again. Besides, for E4D system, it was shown that the conventional grading by the faculty was not located within an acceptable range derived using this software.[17] In another study, grades generated from a digital grading software (Nissan Fair Grader 100, Nissin Dental Products Inc., Kyoto, Japan) of students' preclinical assignment were compared with four calibrated faculty members. This software is able to scan a student's prepared tooth and compares it to a known (software-driven) standardized preparations, and it was revealed that the agreement between scores produced by the digital system and the scores of the examiners was mostly in the low-to-moderate range.[18] Finally, in another study,[19] feasibility of a tooth preparation assessment software referred as Preppr was evaluated as an assistant and educational tool for dental students for comparison of their crown preparation parameters with standard and optimal parameters.

The proposed method presents a device to calculate an immediate and accurate measurement of tooth preparation for dentistry students. A simple portable device with reasonable price and reliable measurement can play a vital role in improving the students' skills and helping them to achieve proper levels of skills during their academic education. In this study, we introduce and evaluate the efficacy of a new and innovative experimental method, including a piece of software and its prerequisite instruments, using image processing techniques in order to remove the role of the faculty in assessment of students' work in preclinical tooth preparation.

  Materials and Methods Top

This study is based on the results of a technology-based research approved by the research council of Isfahan University of Medical Sciences, and the invention is approved by the Intellectual Property Center of State Organization for Registration of Deeds and Properties, Iran (No 98352). The block diagram of the whole process is shown in [Figure 1].
Figure 1: Block diagram of the proposed method

Click here to view

In this research, a piece of software is designed and implemented in Qt platform with C programming language and open CV image processing libraries. Qt is a technology strategy, which lets that programmers quickly and cost-effectively design, develop, deploy, and maintain software while delivering a seamless user experience across all devices. The software has a graphical user interface in which tooth images can be easily loaded and crown preparation parameters are computed and compared with parameters achieved from gold standard images. The source code of this software will be publicly available after the publication of the study. To make use of this software, a prerequisite instrument is designed to hold the sample tooth in a determined location and the imaging is simply done by mobile cell phones. The designed instrument is compatible with almost any cell phone with different sizes and different location of the camera due to flexible parts.

In other software packages such as E4D,[16] the crown prepared by students is scanned with 3D methods and then compared with 3D scans of the crown prepared by individual faculty members as gold standard cases. However, in proposed method, the parameters of crown preparation are computed automatically by image processing techniques on images acquired with mobile cell phones, with no need to complicated 3D scanners. The extracted parameters can be compared with standard and predefined parameters to evaluate the crown preparation qualitatively and quantitatively. This simple and efficient scheme is the main novelty of the proposed software compared to available methods.

Image registration

In the first step, for each tooth, two sets of images are taken from the artificial teeth before and after crown preparation. Photographs were taken using an instrument (box) prepared for this purpose. Camera positions are set in the instrument to make images in 0° and 90° with respect to the horizontal surface (referred as facial and occlusal images, respectively). For this purpose, two adjustable holes, above and in the proximal side, are made, which enables the users to take two images from the facial and occlusal sides. Adjustable holders are designed to adapt to different cell phones. Inside of the box is colored in black, and a set of light sources are spread around the box room to provide the best image quality with the least possible shadow effect. Furthermore, a special place is added to the box room to settle the plastic arc, and a graded movable platform is designed to enable the students to re-settle the arc in its exact place to take fully registered post images.

In this study, the images before and after crown preparation are called theFirstImg and SecondImg, respectively [Figure 2]. Four images are then stacked to calculate crown preparation parameters. The first step in parameter calculation is image registration, which is the process of transforming different sets of images into one coordinate system. Here, two procedures are examined for image registration.
Figure 2: (a) First image: Facial image before crown preparation, (b) second image: Facial image after crown preparation

Click here to view

The first method is a semi-automated method with user interaction. In this method, four points are selected by the user from theFirstImg, followed by selecting four corresponding points from the SecondImg. Then, a transform is found by homographic functions to map the first four points to the second ones. The presence of eight anonymous parameters in homographic functions justifies the need for four points.

The second method is an automated method in which two sets of key points from theFirstImg and SecondImg are detected by a surf detector[20] and then are described by local descriptors. In the next step, the best match between the points among the two sets is found and a transform is computed to provide a registration for the matched points. The registration result ofFirstImg and SecondImg is shown in [Figure 3].
Figure 3: Facial image after registration of FirstImg and SecondImg

Click here to view

Extraction of region of interest

In this specific application, region of interest (ROI) is a limited region located around the prepared tooth. The calculations showing the quality of preparation become very difficult if the region is not limited around the tooth of interest. Therefore, ROI is selected by the user interaction by mouse drag, resulting in a rectangular ROI. [Figure 4] shows a sample occlusal image before and after the ROI selection.
Figure 4: (a) Original occlusal image, (b) Selected region of interest

Click here to view

Tooth segmentation

The main goal of this research is to localize the tooth boundary and compute crown preparation parameters. For this purpose, the following steps are applied sequentially.

Transform to hue, saturation, and value color space

To discriminate the background region from the ROI, color differences are utilized. For this purpose, hue, saturation, and value (HSV) color domain provides a better representation than red, green, and blue (RGB) color coding.[21] In the first step, the reference ranges of H, S, and V components are considered for tooth regions empirically. Then, the HSV components of each pixel are compared with the reference ranges to be assigned to the tooth or background. [Figure 5] demonstrates a binary image after thresholding in HSV color domain.
Figure 5: A binary occlusal image achieved from thresholding in hue, saturation, and value color domain

Click here to view

Edge detection

To compute the tooth perimeter and area before and after tooth preparation, tooth edges should be segmented to lead to tooth contours. Here, the Canny edge detector is used with two thresholds T1 and T2.[22] If the pixel gradient is >T2, the pixel is marked as an edge, and for the case of being between T1 and T2, it is also considered an edge if the pixel is placed close to a current edge. [Figure 6] shows the results of Canny edge detection applied to the selected ROI.
Figure 6: The result of Canny edge detector on selected region of interest from occlusal image in Figure 5

Click here to view

Morphologic operations

As indicated in [Figure 6], extra edge pixels in the background are segmented by the edge detection method. This problem affects the following steps of processing. To address this issue, morphology operations are used.[23] The main purpose of using such operations is to achieve convex and continuous contours. “Erosion” morphology operation is applied to images after the edge detection phase. This operation removes small white regions and creates continuous dark regions in the background. [Figure 7] depicts the segmented edges of an ROI image with and without morphology operation.
Figure 7: Effect of erosion operation. (a) Sample edge detected image before and (b) after erosion

Click here to view

For SecondImgs taken from the occlusal view, two contours are needed to be detected (the internal contour is the resulted contour after teeth preparation). In such case, “dilation” operation is applied after edge detection to find the internal counter. This operation connects the white regions. A sample image is demonstrated in [Figure 8] to show the edges before and after the dilation operation.
Figure 8: Effect of dilation operation. (a) Before and (b) after dilation

Click here to view

Contour extraction

For occlusal images, the interior and exterior contours of the teeth are extracted. To find the exterior contour, different window sizes are considered for the structure element in morphology operations, which lead to extraction of different contours. The concaveness of the extracted contours is compared with a threshold, and a contour with the concaveness more than a threshold is selected. In cases that there is no contour with this condition, the contour with the lowest concaveness is selected. To find the interior contour, edges are detected by the Canny algorithm, followed by the dilation operation to connect discrete edges. The second maximum is selected among all extracted contours. If the concaveness of the selected contour is less than a threshold, it is selected as the final interior contour; otherwise, the contour with the minimum concaveness is selected. For facial images, the tooth perimeter should be considered by a similar method with occlusal images. The samples of the contours by proposed method and the ground truth by experts are depicted in [Figure 9] and [Figure 10].
Figure 9: Contour segmentation in occlusal image. (a) Segmented contours by the proposed method, (b) exterior contour segmented by expert, and (c) interior contour segmented by expert

Click here to view
Figure 10: Contour segmentation in facial image. (a) Segmented contours by the proposed method, (b) segmented contour by expert

Click here to view

Computation of occlusal parameters (distances) for crown Preparation

The main objective of tooth image processing and contour extraction is to determine the irregular and regular parts of tooth preparation. After contour segmentation, the goodness of crown preparation can be quantitatively evaluated. For this purpose, all corresponding points located in the interior and exterior contours are considered counterclockwise [Figure 11].
Figure 11: Corresponding points located in interior and exterior contours

Click here to view

It is expected that two interior and exterior contours have smooth changes and are in a similar distance. Namely, by assigning a threshold, contour differences more than a threshold are considered as an error [Figure 12], red colors].
Figure 12: Crown preparation with differences higher than threshold

Click here to view

Computation of facial parameters (angles) for crown preparation

Crown preparation angle is an important parameter which can be extracted from the facial images. Two lines on two sides of tooth region (after preparation) are considered to compute the crown preparation angle. To this end, first, the center of contours is determined, and then, the contour is divided into right and left sections. Two angles are then calculated for each section, which are expected to be in acceptable range and also to be symmetric enough [Figure 13].
Figure 13: Crown preparation angle in facial image

Click here to view

  Results Top

In this section, we discuss the proposed methods for parameter selection and the corresponding results. In the first step, HSV color space is used for tooth segmentation. To find the tooth color range, in each HSV color component, either the upper bound or lower bound is fixed, and then, another bound is changed to achieve the maximum accuracy for tooth segmentation. For the H component, the upper bound is fixed with 255 and the lower bound is selected among the quantities 5, 10, 15, and 20. Note that in this experiment, all tooth images are segmented correctly only in the lower bounds (5, 10, 15, and 20). [Table 1] presents the accuracy of tooth segmentation in each tooth for different lower bounds of H. It is clear that the lower bound 10 (h_min = 10) achieves the best accuracy (0.89). Therefore, bound range from 10 to 255 is selected as the final range for component H. With a similar procedure, bound ranges from 0 to 93 and from 208 to 255 are selected for the components S and V, respectively. For components H and V, the upper bound 255 is fixed, whereas for component S, the lower bound 0 is fixed.
Table 1: Accuracy of tooth segmentation for occlusal images with selection of different lower bounds for component H

Click here to view

Erosion as a morphology operation is applied to images after the edge detection phase to remove small white regions and to create continuous dark regions in the background. The most important parameter in this operation is the window size, referred to as stride. The larger is the quantity of window size, the bigger white area around the tooth edge is removed (which may lead to elimination of correct tooth edge). For the window size, the upper bound is fixed to 40 and the lower bound is selected among the set 5, 10, 15, 20, 25, 30, 35. The accuracies for different window sizes are reported in [Table 2]. The window size 5 achieves the best average accuracy (0.89).
Table 2: Accuracy of tooth segmentation for occlusal images with selection of different window sizes for erosion operation

Click here to view

The window size for dilation is also explored for the occlusal and facial images. For dilation, the upper and lower bounds are considered as 12 and 5, respectively, with quantities 7 and 10 between them. Note that for window sizes 12 and less, all contours can be segmented correctly. This is the reason for selecting 12 as the upper bound. As it is clear from the reported accuracies for different window sizes in [Table 3], window size of 7 is the best parameter for dilation.
Table 3: Accuracy of tooth segmentation for occlusal images with selection of different window sizes for dilation

Click here to view

The last parameter is the threshold value for the Canny edge detector. Canny uses two thresholds T_min and T_max for edge detection. Here, T_min is considered as 0 and T_max is selected from a range with a similar manner for the erosion and dilation parameters. Note that based on this experiment, Canny can correctly segment the edge of all images only in the interval 20–25 for T_max. Therefore, T_max is evaluated for quantities 20, 23, and 25, in which the highest accuracy (0.88) is observed in T_max = 25 [Table 4].
Table 4: Accuracy of segmentation for occlusal images with selection of different thresholds T_max for edge detection

Click here to view

To show the effectiveness of the proposed method, the results are compared with ground truth labels by experts. The contours segmented by the proposed method and by experts in sample pairs of facial images are shown with cyan and green colors, respectively [Figure 14]. The segmentation results for sample pairs of occlusal images are also depicted in [Figure 15].
Figure 14: Segmented contours of four sample facial images by: (a) Expert and (b) the proposed method

Click here to view
Figure 15: Segmented contours by the proposed method (left image) and expert (right image) in eight sample occlusal images

Click here to view

To evaluate the performance of the proposed method in calculation of facial parameters (angles) for crown preparation, [Table 5] reports the angle computed by the ground truth and the proposed method as well as the difference (signed error) and absolute difference (absolute error) between these angles for 12 images (images 10–21). The best correlation between the crown preparation angle in gold standard and the proposed method is achieved in image 10 with an absolute error of 2.5°, while image 21 is the worst case with an absolute error of 10.85. The mean absolute error for all images is 7.17.
Table 5: The angle of crown preparation computed by the proposed method and experts

Click here to view

The segmentation of desired region in facial images is also evaluated by the Intersection over Union (IOU) criterion. [Table 6] presents the IOU for all images (images 1–21). The best and worst performances are achieved in images 10 and 21 with IOUs of 0.76 and 0.94, respectively. The segmentation results of the proposed method have average an IOU of 0.89 in all images.
Table 6: Intersection over Union for segmentation

Click here to view

The occlusal parameters (distances) for crown preparation is another important output parameter of the proposed method. The internal and external contour segmentations are computed, and the IOU between automatically segmented regions with ground truth is reported in [Table 7]. It is concluded that the external contour with an averaged IUO of 0.93 is segmented more accurately than the internal contour with an averaged IUO of 0.88.
Table 7: Intersection over Union for internal and external contour segmentation

Click here to view

  Discussion Top

Accurate and fair assessment of student works is usually considered as the most critical phase of education. In preclinical dental education, students need to receive consistent and accurate feedback from the faculty in order to achieve a higher level of performance before advancing to the clinics.[24] However, according to the attributors leading to disagreement over student works, including grading scale, rater calibration, and subjective influences, providing consistent feedback by the faculty is very difficult.[25] Valid and objective criteria would help resolve the grading crisis in dental education, which can be partly achieved by removing the human element from evaluation and developing objective evaluation methods.[26],[27]

In our study, we have introduced a novel and convenient method for evaluating the quality of metal-ceramic tooth preparation by dental students. However, the innovated instrument has the potential to be planned for other crown materials by defining the cutoff points. This invention may improve the objective and repeatable judgment of students' work in preclinical settings. Moreover, this method does not require high-tech technologies such as scanner systems and is more user-friendly. The pre- and postpreparation photos can be taken by students' cellphones.

In the proposed method, the best results in segmentation of contours are achieved in cases with good imaging quality (appropriate light and imaging angle). However, our segmentation method is designed to be less sensitive to changes of such conditions. Furthermore, perfect segmentation is evident in specific regions of the image with an appropriate range of RGB color spaces. As shown in [Figure 13], in some cases where these conditions have significant changes, the proposed method fails in correct segmentation. For example, if the color of gingiva is close to the dental color, the proposed method misleads the contour path and segments a smaller dental region. This scenario is also valid for both facial and occlusal images. For internal contours, when the edge of crown preparation is not clear, the contour will be segmented incorrectly and leads to a lower accuracy. Furthermore, if the side regions of the teeth are changed significantly, it is difficult to fit a single line and an error in angle computation appears.

  Conclusions Top

In this research, a new package, including software and its prerequisite instruments, was proposed to help students in preclinical tooth preparation. For this purpose, images were taken from artificial teeth in different angles and directions before and after crown preparation and then were analyzed by the proposed methods to evaluate the quality and quantity of crown preparation. The proposed tooth image processing method had image registration, ROI segmentation, transformation to HSV color space, edge detection, morphology operation, and contour extraction abilities, which computed the amount and angle of crown preparation. Applying the proposed platform to a local tooth dataset showed a mean absolute error of 7.17 and a mean IOU of 0.89 in comparison to gold standard images segmented by the experts. Our method of evaluation seems totally a convenient aid to dental students to evaluate their work. However, before the application of this method as a part of dental education, the accuracy and validity of the outcomes should be tested against the faculty grading, a situation happening routinely in dental schools.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.

  Biographies Top

Bahareh Tahani received the D.D.S degree in General Dentistry from Isfahan University of Medical Sciences, Iran, in 2003, and PhD degree in Community Dentistry from Shahid Beheshti University of Medical Sciences, Iran, in 2013. Her field of interest is health system and dental service research. She has conducted some national and regional research in the field of Health Technology Assessment. From 2013, she has joined Isfahan University of Medical Sciences as Assistant Professor at the Department of Oral Public Health.

Email: [email protected]

Abdolreza Rashno received the B.S. degree in computer engineering from Shahid Chamran University of Ahvaz, Iran, in 2009, M.S. degree (with honors) in artificial intelligence from Kharazmi University of Tehran, Iran, in 2011 and Ph.D. degree (with the best thesis national award) in artificial intelligence from Isfahan University of Technology, Iran in 2018. He was an outstanding researcher at University of Minnesota, USA, 2015-2017. From 2018, he joined Lorestan University as Assistant Professor at the Department of Computer Engineering. His research interests include image segmentation, medical image analysis, image restoration, content-based image retrieval, evolutionary computing and machine learning.

Email: [email protected]

Hamed Haghighi i received my MSc in Artificial Intelligence from the University of Tehran in 2019 and my BSc in Software Engineering from the Isfahan University of Technology in 2016. My main research interests are generative models, deep learning, computer vision, and audio signal processing. In detail, I work on modeling dependencies in high fidelity images and audio using deep generative models. While I mainly focus on these areas, I am also interested in other domains such as natural language processing, computer graphics, and any other artificialintelligence- related task.

Email: [email protected]

Rasoul Monirifard received the D.D.S degree in General Dentistry from Isfahan University of Medical Sciences, Iran, in 2003, and Msc in Prosthodontics from Isfahan University of Medical Sciences, Iran, in 2008. His field of interest is fix prosthodontics . He has conducted several researches in that field. He worked as assistant professor in department of prosthodontics between 2008 and 2015 and in private office by now.

Email: [email protected]

Hooman Nosrati Khomami received his Masters degree in Biomedical engineering from Sepahan Olum and fanavari. His bachelor degree was also obtained in Power Electronics from Nagsh e Jahan University of Isfahan.

Email: [email protected]

Rahele Kafieh is a guest researcher at Neurocure Clinical Research Center, Charite University, Berlin, Germany. She already work on deep learning for segmentation and classification of ocular data in Neurodegenerative diseases. She is also an Assistant Professor at School of Advanced Technologies in Medicine, Isfahan University of Medical Sciences, Isfahan, Iran. Her research is concentrated on biomedical image analysis. She is interested in problems in area of graph based image analysis, timefrequency methods, deep learning, image segmentation, restoration and others.

Email: [email protected]

  References Top

Goodacre CJ, Bernal G, Rungcharassaeng K, Kan JY. Clinical complications in fixed prosthodontics. J Prosthet Dent 2003;90:31-41.  Back to cited text no. 1
Shillingburg HT, Hobo S, Whitsett LD, Jacobi R, Brackett S. Fundamentals of Fixed Prosthodontics. Vol. 194. Jordan University of Science and Technology Faculty of Applied Medical Sciences Department of Applied Dental Sciences: Quintessence Publishing Company; 1997.  Back to cited text no. 2
Poggio CE, Ercoli C, Rispoli L, Maiorana C, Esposito M. Metal-free materials for fixed prosthodontic restorations. Cochrane Database Syst Rev 2017;12:CD009606.  Back to cited text no. 3
Pjetursson BE, Sailer I, Zwahlen M, Hämmerle CH. A systematic review of the survival and complication rates of all-ceramic and metal-ceramic reconstructions after an observation period of at least 3 years. Part I: Single crowns. Clin Oral Implants Res 2007;18 Suppl 3:73-85.  Back to cited text no. 4
. McLean JW, “The Science and Art of Dental Ceramics,” The Nature of Dental Ceramics and their Clinical Use, 1979: p. 79-82.  Back to cited text no. 5
Rosenstiel SF, Land MF. Contemporary Fixed Prosthodontics-E-Book. Future university of egypt: Elsevier Health Sciences; 2005.  Back to cited text no. 6
Dodge W, Weed RM, Baez RJ, Buchanan RN. The correlation of resistance and retention to convergence angle. In journal of dental research. Am Assoc Dent Res 1983;62:267.  Back to cited text no. 7
Hovijitra S, Robinson F, Brehm T. The relationship between retention and convergence of full crowns when used as fixed partial denture retainers. J Indian Dent Assoc 1979;58:21-4.  Back to cited text no. 8
Mack PJ. A theoretical and clinical investigation into the taper achieved on crown and inlay preparations. J Oral Rehabil 1980;7:255-65.  Back to cited text no. 9
Goodacre CJ. Palladium-silver alloys: A review of the literature. J Prosthetic Dent 1989;62:34-7.  Back to cited text no. 10
Al Amri MD, Sherfudhin HR, Habib SR. Effects of Evaluator's fatigue and level of expertise on the global and analytical evaluation of preclinical tooth preparation. J Prosthodont 2018;27:636-43.  Back to cited text no. 11
Liu L, Li J, Yuan S, Wang T, Chu F, Lu X, et al. Evaluating the effectiveness of a preclinical practice of tooth preparation using digital training system: A randomised controlled trial. Eur J Dent Educ 2018;22:e679-86.  Back to cited text no. 12
Sirona. Sirona prepCheck. Computer-Aided Training System: A Plus for Universities and Students; 2019. Available from: www.sirona.com/en/products/treatment-centers/prepcheck. [Last accessed on 2019 June 05].  Back to cited text no. 13
Kunkel TC, Engelmeier RL, Shah NH. A comparison of crown preparation grading via PrepCheck versus grading by dental school instructors. Int J Comput Dent 2018;21:305-11.  Back to cited text no. 14
Schepke U, van Wulfften Palthe ME, Meisberger EW, Kerdijk W, Cune MS, Blok B, et al. Digital assessment of a retentive full crown preparation – An evaluation of prepCheck in an undergraduate preclinical teaching environment. Eur J Dent Educ 2020;00:1-18. doi: 10.1111/eje.12516.  Back to cited text no. 15
Renne WG, McGill ST, Mennito AS, Wolf BJ, Marlow NM, Shaftman S, et al. E4D compare software: An alternative to faculty grading in dental education. J Dent Educ 2013;77:168-75.  Back to cited text no. 16
Callan RS, Haywood VB, Cooper JR, Furness AR, Looney SW. et al. The validity of using E4D compare's “% comparison” to assess crown preparations in preclinical dental education. J Dent Educ 2015;79:1445-51.  Back to cited text no. 17
Kateeb ET, Kamal MS, Kadamani AM, Abu Hantash RO, Arqoub MM. Utilising an innovative digital software to grade pre-clinical crown preparation exercise. Eur J Dent Educ 2017;21:220-7.  Back to cited text no. 18
Tiu J, Cheng E, Hung TC, Yu CC, Lin T, Schwass D, et al. Effectiveness of crown preparation assessment software as an educational tool in simulation clinic: A pilot study. J Dent Educ 2016;80:1004-11.  Back to cited text no. 19
Bay H, Ess A, Tuytelaars T, Gool LV. Speeded-up robust features (SURF). Comput Vis Image Underst 2008;110:346-59.  Back to cited text no. 20
Chernov V, Alander J, Bochko V. Integer-based accurate conversion between RGB and HSV color spaces. Comput Electr Eng 2015;46:328-37.  Back to cited text no. 21
Shrivakshan G, Chandrasekar C. A comparison of various edge detection techniques used in image processing. Int J Comput Sci Issue (IJCSI) 2012;9:269.  Back to cited text no. 22
Lee J, Haralick R, Shapiro L. Morphologic edge detection. IEEE J Robot Autom 1987;3:142-56.  Back to cited text no. 23
Feil PH, Gatti JJ. Validation of a motor skills performance theory with applications for dental education. J Dent Educ 1993;57:628-33.  Back to cited text no. 24
Fuller JL. The effects of training and criterion models on interjudge reliability. J Dent Educ 1972;36:19.  Back to cited text no. 25
Sharaf AA, Abdel Aziz AM, El Meligy OA. Intra-and inter-examiner variability in evaluating preclinical pediatric dentistry operative procedures. J Dent Educ 2007;71:540-4.  Back to cited text no. 26
Schiff AJ, Salvendy G, Root CM, Ferguson GW, Cunningham PR. Objective evaluation of quality in cavity preparations. J Dent Educ 1975;39:92-6.  Back to cited text no. 27


  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9], [Figure 10], [Figure 11], [Figure 12], [Figure 13], [Figure 14], [Figure 15]

  [Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]


Similar in PUBMED
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

  In this article
   Materials and Me...
   Article Figures
   Article Tables

 Article Access Statistics
    PDF Downloaded64    
    Comments [Add]    

Recommend this journal