|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2014 | Volume
: 4
| Issue : 3 | Page : 115-121 |
|
Identification of pulpitis at dental X-ray periapical radiography based on edge detection, texture description and artificial neural networks
Bernard Y Tumbelaka1, Fahmi Oscandar2, Faisal Nur Baihaki1, Suhardjo Sitam2, Mandojo Rukmo3
1 Department of Physics, Faculty of Mathematics and Natural Sciences, Faculty of Dentistry, University of Padjadjaran, Jatinangor, Indonesia
2 Department of Radiology Dentistry, Faculty of Dentistry, University of Padjadjaran, Jatinangor, Indonesia
3 Department of Conservative Dentistry, Faculty of Dentistry, University of Airlangga, Surabaya, Indonesia
| Date of Web Publication | 6-Aug-2014 |
Correspondence Address:
Bernard Y Tumbelaka
Faculty of Mathematics and Natural Sciences, University of Padjadjaran, Jatinangor
Indonesia
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/1658-5984.138139
Objectives: The aim of the present research was to identify pulpitis through periapical radiography by applying edges as basis image features, the texture description and the artificial neural networks (ANNs).
Materials and Methods: Input image data records of 10 molar and 10 canine teeth were used. The clinical diagnosis of interest cases were represented as normal pulp, reversible and irreversible pulpitis, and necrotic pulp. The following image processing steps were done. First, the data records were converted digitally and preprocessed as its original image using the Gaussian Filter to obtain the best smoothed intensity distribution. Second, the local image differentiation was used to produce edge detector operators, e(x,y) as the image gradient; ∇f(x,y) providing useful information about the local intensity variations. Third, these results were analyzed by using the texture descriptors to obtain digitally the image entropy, H. The fourth step, all were characterized by the ANNs. Results: The edge detection carried important information about the object boundaries of pulpal health and pain conditions in the dental pulp significantly. The image entropy which was identified, the diagnostic term, was obtained from texture descriptors in the segmentation regions where the curves of pulp states tent convergence with the normal pulp line from 4.9014 to 4.6843 decreasing to the reversible and the irreversible pulpitis line include the nectrotic pulp line from 4.6812 to 4.5926 and then inputting to the ANNs analysis at the same of mean square error around 0.0003. Conclusions: Referred to these results, the correlation of the image entropy and the ANNs analysis could be linearly classified with the critical point of 4.6827. Finally, it could be concluded that the direct reading radiography is better to be digitized in order to provide us the best choice for diagnosis validation. Keywords: Artificial neural networks, dental X-ray, edge detection, image entropy, mean square error, periapical radiography, pulpitis, texture description
How to cite this article:
Tumbelaka BY, Oscandar F, Baihaki FN, Sitam S, Rukmo M. Identification of pulpitis at dental X-ray periapical radiography based on edge detection, texture description and artificial neural networks. Saudi Endod J 2014;4:115-21 |
How to cite this URL:
Tumbelaka BY, Oscandar F, Baihaki FN, Sitam S, Rukmo M. Identification of pulpitis at dental X-ray periapical radiography based on edge detection, texture description and artificial neural networks. Saudi Endod J [serial online] 2014 [cited 2023 Mar 21];4:115-21. Available from: https://www.saudiendodj.com/text.asp?2014/4/3/115/138139 |
| Introduction | |  |
It is a fact that there has been almost no one to pay attention earlier until the stage in the decay process is usually too late for preventive and conservative intervention. Currently in Indonesia, dental decay has often been diagnosed using radiographic techniques. [1] Until now it is impossible to detect and monitor the stages of dental decay process of pulpitis by periapical radiography which has the sensitivity for this type of lesion. Eventually the diseased tissue has been strangling the blood vessels of the dental pulp that leads to death. At this point, pulpitis has become the most common cause of toothache followed by death of the pulp and spread of infection through the apical foramina into the periapical tissues. [2],[3] This could be found among all ages especially younger people due to several unhygienic factors like smoke, hormonal disturbance in woman, diabetes, stress, cancer, AIDS, genetic factors, insufficiently nutrients, and drugs. [4]
An attempt to find a new method in radiography that could identify this inflammation of tooth pulp in soft tissue is worth to try it. The pulp is the inner part of the tooth that consists of blood vessels, nerve endings, and connective tissues. The primary objective of pulp therapy is to maintain the integrity and health of the teeth and their supporting tissues. The indications are based on the clinical diagnosis of normal pulp, reversible pulpitis (pulp is capable of healing), symptomatic or asymptomatic irreversible pulpitis (vital inflamed pulp is incapable of healing) and nectrotic pulp or pulpless (pulp do not respond to vitality test). [5],[6],[7],[8] It might be interesting to try to find these indications that could be obtained from radiographic evidence of periapical radiolucency to diagnose pulpitis. It was impossible to achieve an accurate diagnosis of the state of the pulp on the basis of clinical proof alone. The only accurate method is to do a histological examination. [9] Therefore, numerous classifications of pulp disease had only identified by a limited number of clinical diagnostic tests before effective dental treatment is performed. Therefore, our research interest was aimed to identify pulpitis in the regions of interest through periapical radiography by applying edges as basis image features, the texture description and the artificial neural networks (ANNs).
| Materials And Methods | | |
Input image data of 10 molar and 10 canine teeth of periapical radiography of X-ray were used. The clinical diagnosis of interested cases was separated into four tooth conditions of pulpal diagnose states: normal pulp, reversible and irreversible pulpitis and necrotic pulp. The following image processing steps were done: The first step, the data records were converted digitally and preprocessed as its original image using the Gaussian filter [10],[11] to obtain the best smoothed intensity distribution. The second step, the local image differentiation technique was used that can produce edge detector operators, e(x,y) as the image gradient; ∇f(x, y) providing useful information about the local intensity variations. The third step, these results were analyzed by using the texture descriptors based on mean and variance analysis that could be used in order to obtain digitally the image entropy, H. [12],[13],[14] The fourth step, all was characterized by the ANNs. [15] We can separate the pulpitis. It could be also separated the infected line in two regions of interest as the reversible pulpitis and the irreversible pulpitis. The process used weight input generated that was always trained until reached the result of learning appropriate data same as identified. For the first training as compared variables let amount of database. Furthermore, these selected number of database input would be very influence the analysis results because the identification error would be able corrected if only if the weight input database were more added.
Experimental setup
These research works were arranged for the preprocessing and processing of the original image to obtain the best intensity distribution using Gaussian filter and to identify the tooth in the next step by the edge detection as the basic image features of the disinfected and infected tooth regions of image segmentation separately to find the deterministic processing [Figure 1]. When the non-deterministic cases were obtained, it was needed to apply the ANNS analysis. | Figure 1: Preprocessing and processing of the original image using Gaussian filter and the edge detection as the basic image features of the disinfected and infected tooth regions of image segmentation
Click here to view |
Gaussian filter
It was needed to apply a very commonly used filter, which smoothly, reduces noise and gets small details. Equation [1] of a Gaussian function in two dimensions was the product of two of one-dimensional Gaussians, one in each dimension, Gσ (x, y) as follows:
where: x is the distance from the origin in the horizontal axis, y is the distance from the origin in the vertical axis, σ is the standard deviation of the Gaussian distribution.
It was also needed to use the discrete convolution of equation [2] as follows:
where: f is the image function, g is the Gaussian function
When it was applied to the edge detection, it might be a derivative of a noisy signal that could cause more noise. Therefore, it was needed to apply the Gaussian filter before taking a smoothed derivative. The differentiation and convolution of equation [3] both linear operators where they commuted together as
Edge detection
Image edges were the form of local variations of image intensity that can produce local image differentiation techniques defined as edge detector operators, the image gradient, ∇f(x,y) could be written as equation [4]:
The local intensity direction could be described by the direction angle of the phase of f(x,y) as follows in equation [7]:
Texture descriptor analysis was based on the edge detection features related to the image entropy.
Texture descriptors
Mostly, an image characteristic depend on its texture used in the region segmentation. There were several important simple texture descriptors such as the image histogram p f (f k ), the arithmetic mean, μ in equation [8] and the variance of the standard deviation square, σ2 in equation [9] and the image entropy, H in equation [10] as a scalar value representing the entropy of gray scale image, f k having B pixels are given by:
where: p f (f k ) - the image histogram, f k - is the various image intensity levels, 0 ≤ f ≤ 2 B , k - is the pixel location, k = 1, 2,…., B, B - gray scale unit, 2 bits, 4 bits,….
Usually, entropy was a statistical measure of randomness that can be used to characterize the texture of the input image histogram. The image histogram was calculated within an image region, f.
The relation between the average codeword length, L(f), and the image entropy was very close that could be written as equation [11]:
Entropy of the gray scale image contained of thousand bits of information, representation of intensity. Entropy converted any class other than logical to unit 8 for the histogram count calculation so that the pixel values were discrete and directly correspond to a bin value. Using these texture descriptors, it was deterministic analysis that could refuse several descriptor cause the reversible and the irreversible problems both the disinfected and the infected tooth regions. Therefore, it was needed to use the ANNs analysis to be the non-deterministic (random) analysis.
Artificial neural networks
Pulpitis identification was done through ANNs analysis, where accuracy was the percentage of number of true classified compared to all final
result classified using mean square error (MSE)
defined as follows in equation [12]:
where: N − number of database, i = 1, 2, 3., N, T − final data output, O - true data output.
| Results | | |
Results are summarized in [Table 1] and [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]. Ten tooth radiographs were used according to the ROI (region of interest) of 10 cropped samples for normal pulp, reversible and irreversible pulpitis, and necrotic pulp. There could be several results described as follow. | Figure 2: Original Images, Gaussian filter images, cropping and edge detection at normal pulp (a) reversible and irreversible pulpitis (b) tooth regions of image segmentations
Click here to view |
 | Figure 3: Curves of variance and mean analysis for normal pulp, reversible, irreversible, necrotic
Click here to view |
 | Figure 4: Correlation between ROI of f and ROI of disinfected and infected by pulpitis based mean and variance
Click here to view |
 | Figure 5: Curve of image entropy of normal pulp, reversible and irreversible pulpitis, and necrotic pulp
Click here to view |
 | Figure 6: Diagnose the best training performance in 32 × 32 iterated at epoch 7 of data testing simulation using by ANNs analysis
Click here to view |
 | Table 1: The relationship between variance and mean regions for normal pulp, reversible and irreversible pulpitis, necrotic pulp regions
Click here to view |
First, the edge detection carried important information about the object boundaries as normal pulp, reversible and irreversible pulpitis, and necrotic pulp significantly which could be valuable for the pulp interpretation as shown in [Figure 2]a and b.
Second, as shown in [Figure 3] and [Figure 4] and [Table 1], first by using mean analysis was obtained directly for normal pulp of region 1 (171.3669) to region 8 (171.5720), and normal pulp to pulptis line separated by reversible of region 8 to region 12 (186.7028), irreversible of region 12 to region 17 (185.5005). Second, variance analysis was also used for normal pulp of region 1 (62.9592) to region 8 (54.8623), and normal pulp to pulpitis line separated by reversible of region 8 to 12 (55.2579), irreversible of region 12 to 17 (46.5700). In [Figure 5], by using image entropy obtained from texture descriptors by mean and variance analysis [15] as deterministic and random curve. The curves of normal pulp and pulpitis regions were figured convergence with normal pulp line from 4.9014 to 4.6843 decreased to pulpitis line from 4.6812 to 4.5926 with MSE around 0.0003, as shown in [Figure 6]. The results could be inputting and expanding our observation figured convergence at the normal pulp and pulpitis regions with normal pulp line decreased to pulpitis line with the critical point by the texture description and the ANNs analysis at the same of MSE around 0.0003. [Figure 6] gave the nonlinearity of the MSE curve representing the accuracy of final output over the true output during the training. Pulpitis identification was done through ANN analysis, where accuracy was the percentage of number of true classified compared to all final result classified. For observation that resulted with the reaching true accuracy below 80% gave the identification still wrong. Therefore, it was needed to increase the level of true accuracy into limit of 95% that means to have to reach the MSE equal to 0.0003.
The current research had owned the advantage diagnosis to obtain the regions of interest more precisely. We could separate the normal pulp line into two regions of interest as the pure normal pulp (4.9577 to 4.8442) and the impure pulp called the regular pulpitis (4.8442 to 4.6827). We could also separate the pulpitis line into two regions of interest as the reversible pulpitis (4.6827 to 4.6565) and the irreversible pulpitis (4.6565 to 4.3973).
| Discussion | | |
Even though the periapical radiographic film was of high quality, there might be more important information that could not be directly detected in regard to the pulp disease states. The present research involved to reduce the diagnosis suspect and to gain accurate diagnosis that could be used to increase the probability of a successful clinical outcome. The biological states of the pulpal tissues had a tendency to give correlating symptom and radiographic diagnosis. Digital image analysis as part of image technology has been enormously developed in dentistry that aimed to interpret periapical radiograph. [16],[17] In endodontics, the quality of the image is very important as it facilitates accurate interpretation of the morphology of the root and the root canal space before establishing the treatment and the during the treatment as well as postoperative and long-term evaluation of the outcome of endodontic treatment. [18],[19],[20],[21]
The principles for diagnosis of the condition of the dental pulp are no different from those applied to the management of other disease conditions. One of the essential methods is radiographic examination which is an important complement to the clinical examination. It provides information about the presence and severity of caries, deep restorations in relation to the dental pulp as well as root fractures and periapical changes.
Our research started with the dental image quality problem that was usually degraded because of its mechanical, optical and electronic noises from its medical tools were utilized. A Gaussian filter smoothes an image by calculating weighted averages in a filter box. [22] Therefore, we had to perform the image pre-processing using a Gaussian filter approximated to its original image as the best intensity distribution that analyzes its image identification and feature parameters linked with the features on object represented in the image. The image features of edge detection represented and correlated with means and variances identifying the state of normal pulp, pulpitis and necrotic pulp as a deterministic solution. These indicated ambiguous results. A nondeteministic solution was needed using the texture description of the grey level image analysis statistically by the image entropy. The probability results were represented by the image entropy, H(B). It has owned the advantage diagnosis to obtain the regions of interest (ROI) using image entropy H(B) more precisely than diagnoses radiography film directly even though compared with the EPT standard. [23],[24] It could expand our observation at the curves of normal pulp, reversible and irreversible pulpitis, necrotic pulp state regions figured convergence with normal pulp state line decreases to pulpitis line with the critical point by the texture description and the ANNs analysis. It might be interesting to develop a learning process from number of database changed for user's needs with high flexibility and standardization.
Although there were always some exceptions, the different symptoms usually detect that a patient presents with the pulpal diseases. Image entropy could be expanded to multivariate function where the pulp pain in higher sensitivity degenerated to the state of dead pulpal tissue. This will be our new and future research goal.
| Conclusions | | |
The correlation of the image entropy and the ANNs analysis could be linearly classified with the critical point of 4.6827. Finally, it could be concluded that the direct reading radiography is better to be digitized in order to provide us the best choice for diagnosis validation.
| Acknowledgments | | |
We acknowledge the great collaboration between Faculty of Dentistry and Faculty of Mathematics and Natural Sciences of University of Padjadjaran and Faculty of Dentistry, University of Airlangga, Indonesia.
| References | | |
| 1. | Mulyani D, McIntyre J. Caries inhibitory effect of fluoridated sugar in a trial in Indonesia. Aust Dent J 2002;47:314-20. 
|
| 2. | Gutmann JL, Baumgartner JC, Gluskin AH, Hartwell GR, Walton RE. Identify and define all diagnostic terms for Periapical/Periradicular health and disease states. J Endod 2009;35:1658-74.
|
| 3. | Dos Santos Guimarães C, Sampaio GC, Barros de Almeida E, Silva LB, Mourato AP, Alves DF. Influence of Apical Foramen Cleaning on Periapical Lesion Repair. Rev Fac Odontol Porto Alegre 2011;52:45-7.
|
| 4. | Mayo Clinic, Oral health: A window to your overall health, Mayo Foundation for Medical Education and Research. Available from: http://www.mayoclinic.org/dental/[Last accessed on 2013 May 11].
|
| 5. | Abbott P, Yu C. A clinical classification of the status of the pulp and the root canal system. Aust Dent J 2007;52 1 Suppl:S17-31.
|
| 6. | Levin LG, Law AS, Holland GR, Abbott PV, Roda RS. Identify and Define All Diagnostic Terms for Pulpal Health and Disease States. J Endod 2009;35:1645-57.
|
| 7. | Newton CW, Hoen MM, Goodis HE, Johnson BR, McClanahan SB. Identify and determine the metrics, hierarchy, and predictive value of all the parameters and/or methods used during endodontic diagnosis. J Endod 2009;35:1635-44.
|
| 8. | Endodontics: Colleagues for Excellence Newsletter. Endodontic Diagnosis. Ame Assoc Endod. Available from: http//:www.aae.org/colleagues [Last accessed in 2013].
|
| 9. | Seltzer S, Bender IB, Ziont ZM. The dynamics of pulp inflammation: Correlations between diagnostic data and actual histological findings in the pulp. Oral Surg Oral Med Oral Pathol 1963;16:969-77.
|
| 10. | Mathematica, Gaussian Filter function. Available from: http://www.reference.wolfram.com/mathematica/ref/GaussianFilter.htm. [Last accessed on 2012].
|
| 11. | Fisher R, Perkins S, Walker A, Wolfart E. Spatial Filters: Laplacian of Gaussian. Available from: http://www.homepages.inf.ed.ac.uk/rbf/HIPR2/log.htm [Last accessed on 2003].
|
| 12. | Conners RW, Harlow CA. A Theoretical comparison of texture algorithms. IEEE Trans Pattern Anal Mach Intell 1980;2:204-22.
|
| 13. | Pitas I. Digital image processing algorithms. United States: Prentice-Hall, Inc. Upper Saddle River, NJ, USA. 1993
|
| 14. | Gonzalez RC, Woods RE, Eddins SL. Digital image processing using Matlab, NJ. Chap 11. United States: Prentice Hall; 2003.
|
| 15. | Matlab products. Available from: http://www.mathworks.com/matlabcentral/fileexchange/26694-gray-level-run-length-matrix/content/GLRLM.[Last accessed on 2012 Dec 12].
|
| 16. | Nair MK, Nair UP. Digital and Advanced Imaging in Endodontics: A Review. J Endod 2007;33:1-6.
|
| 17. | Abella F, Patel S, Duran-Sindreu F, Mercadé M, Bueno R, Roig M. An evaluation of the periapical status of teeth with necrotic pulps using periapical radiography and cone-beam computed tomography. Int Endod J 2014;47:387-96.
|
| 18. | Nance R, Tyndall D, Levin LG, Trope M. Identification of root canals in molars by tuned-aperture computed tomography. Int Endod J 2000;33:392-6.
|
| 19. | Lozano A, Forner L, Llena C. In vitro comparison of root-canal measurements with conventional and digital radiology. Int Endod J 2002;35:542-50.
|
| 20. | Jung M, Lommel D, Klimek J. The imaging of root canal obturation using micro-CT. Int Endod J 2005;38:617-26.
|
| 21. | Yoshioka T, Kobayashi C, Suda H, Sasaki T. An observation of the healing process of periapical lesions by digital subtraction radiography. J Endod 2002;28:589-91.
|
| 22. | Liu Ce, Szeliski R, Kang SB, Zitnick CL, Freeman WT. Automatic Estimation and Removal of Noise from a Single Image. IEEE Trans Pattern Anal Mach Intell 2008;30:299-314.
|
| 23. | McClanahan SB, Johnson JD, Blundell RE Jr. Pulpal and periradicular diagnostic terminology, Naval Postgrad Dental School National Naval Dental Center, Bethesda, Maryland, Clinical Update, Vol. 24, No. 1 January 2002.
|
| 24. | American Academy of Pediatric Dentistry (AAPD), Guideline on Pulp Therapy for Primary and Immature Permanent Teeth, Reference Manual, IV 34, No 6, 12/13, 2009. p. 222-9.
|
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]
[Table 1]
| This article has been cited by | | 1 |
Developments and Performance of Artificial Intelligence Models Designed for Application in Endodontics: A Systematic Review |
|
| Sanjeev B. Khanagar, Abdulmohsen Alfadley, Khalid Alfouzan, Mohammed Awawdeh, Ali Alaqla, Ahmed Jamleh | | Diagnostics. 2023; 13(3): 414 | | [Pubmed] | [DOI] | | | 2 |
Research on Edge Detection Algorithm of Indoor Soft Decoration Pattern |
|
| Hao Zheng, Lixia Wang, Jiguo Yu | | Mobile Information Systems. 2022; 2022: 1 | | [Pubmed] | [DOI] | | | 3 |
Dental disease detection on periapical radiographs based on deep convolutional neural networks |
|
| Hu Chen,Hong Li,Yijiao Zhao,Jianjiang Zhao,Yong Wang | | International Journal of Computer Assisted Radiology and Surgery. 2021; | | [Pubmed] | [DOI] | | | 4 |
Selection of an optimal treatment method for acute pulpitis disease |
|
| Bahadur F. Aliev,Latafat A. Gardashova | | Procedia Computer Science. 2017; 120: 539 | | [Pubmed] | [DOI] | |
|
 |
|
|
|
Search Pubmed for