Determination of Abcession Size Using Digital Image Processing to Aide Endodontic Therapy

Only available on StudyMode
  • Topic: Root canal, Endodontic therapy, Digital image processing
  • Pages : 5 (1660 words )
  • Download(s) : 117
  • Published : January 11, 2013
Open Document
Text Preview
DETERMINATION OF ABCESSION SIZE USING DIGITAL IMAGE PROCESSING TO AIDE ENDODONTIC THERAPY Karthikeyan Karunakaran, Nikil Ravi, Srinevasan MS

Abstract-- Endodontic therapy is a treatment where the infected tooth s cavity is removed completely and filled with gutta-percha .During this therapy the size of the abscission in the root apex is important. If the abscission size is small the dentists call them lesion which can be easily removed during the therapy itself. But if it is large, the therapy won’t be effective & will cause acute pain to the patient .In such a way the abscession must be removed by injecting penicillin for a week. But current techniques cannot discriminate between a lesion and an abscess .Our paper aims at distinguishing these two by astute measurement of their respective sizes using two algorithms –hill climbing algorithm, fuzzy C mean clustering algorithm aided by fundus photograph. Keywords: Abscission, Fuzzy C, Hill climbing, Endodontic therapy, Lesion, Gutta- percha. I INTRODUCTION

Human teeth will be more prone to infection if it is void of any care. In the recent past many people are plagued by dental infections. Endodontic therapy (root canal) is used in the elimination of the infection in contaminated tooth. Each tooth has a pulp chamber which usually consists of nerve tissues, blood vessels and cellular entities .Endodontic therapy involves discarding of above mentioned structures, cleaning and filling of the pulp chamber. Gutta –percha (an eugenol base cement). After the therapy the tooth is theoretically dead since it is void of any stimulus. It becomes brittle so the tooth is fitted with an external crown. But there is another condition which should be checked before performing root canal therapy. Every tooth has a swelling in its root apex. If the swelling is of negligible size, it is referred as lesion in dental terms. If it’s a lesion then no other additional pre-treatment is necessary. But if the swelling is large, then it’s a serious case to consider .If no pre-treatment is performed then the patient will encounter acute pain which dilutes root canal‘s advantage. Present methods fail to aid the dentists with accurate information about the size of a lesion and an abscess .So the dentist acts blindfold and performs root canal which more often than not ends with the patient screaming with pain. In our proposed idea we can provide accurate differences between the lesion and abscession by using two different algorithms in digital image processing .They are hill climbing algorithm and fuzzy C mean clustering algorithm. II ALGORITHMS

In this section we are going to extensively look about the two algorithms. A.HILL CLIMBING ALGORITHM:
In this paper we are using hill climbing algorithm because it can provide accurate discrimination between the abscess and the root apex which is also on the same axis. Other algorithms such as watershed algorithms fail in this front. We use hill climbing algorithm to determine the region of the swell and the second algorithm to measure the size. Before processing the image we should de-noise it. B. DENOISE:

The fundus image of the tooth is taken .De noising is achieved by the mechanism of anisotropic diffusion. We use this particular mechanism because information of the previous noise pattern is not needed. It provides better contrast .Here the noises are blurred and the edges are preserved through the diffusion equation combined with the edge information . I t = div (c(*x,y,t)grad I)=c(x,y,t)grad I+(grad c).(grad I) Where div is the divergence operator, grad is a gradient and c is the conduction coefficient function, index t denotes the time.

C.DETERMINATION OF APPROXIMATE REGION:
In our paper the regions of interest (RoI) is the abscess. Actually there are three planes in an RGB image –G plane, R plane, B plane. Our input image is an RGB image. Now to identify the approximate region we use G plane because it provides better contrast. The most...
tracking img