Ekoputris, Rizqi Okta (2017) Segmentasi Trabecular Bone Dental Panoramic Radiograph Berbasis Karakteristik Profil Segmen Menggunakan Extreme Learning Machine. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
5113100005-Undergraduate_Theses.pdf - Published Version Download (2MB) | Preview |
Abstract
Dental panoramic radiograph adalah citra x-ray dua dimensi (2-D) dari gigi yang merekam keseluruhan mulut, termasuk gigi, rahang atas, rahang bawah dan jaringan serta struktur yang melingkupinya dalam satu citra. Pada dental panoramic radiograph mengandung banyak informasi yang dapat diidentifikasi salah satunya melalui struktur trabecular bone.
Penelitian ini mengusulkan segmentasi area trabecular bone pada dental panoramic radiograph berbasis karakteristik profil segmen dan metode klasifikasi Extreme Learning Machine. Input dari metode ini adalah dental panoramic radiograph. Pemilihan region of interest (ROI) dilakukan pada tulang rahang bawah pada area trabecular bone yang didalamnya terdapat gigi dan cortical bone. ROI tersebut dibagi lagi menjadi dua dimana ROI atas mengandung gigi dan ROI bawah mengandung cortical bone. Setelah itu, hasil pemotongan ROI dilakukan preprocessing menggunakan filter mean dan median untuk ROI atas dan filter motion blur untuk ROI bawah. Citra yang telah terpisah tersebut masing-masing diekstrak tiap pikselnya menjadi 4 fitur yang terdiri dari intensitas citra, filter Gaussian 2D dengan dua sigma yang berbeda, dan filter Log Gabor untuk ROI atas. Untuk ROI bawah, digunakan 5 ekstraksi fitur yaitu intensitas citra, filter Gaussian 2D dengan dua sigma yang berbeda, phase congruency, dan Laplacian of Gaussian. Kemudian digunakan beberapa sampel piksel sebagai data training untuk membuat model Extreme Learning Machine. Output dari classifier ini adalah area segmentasi dari trabecular bone.
Pada ROI atas, didapatkan rata-rata sensitivitas, spesifisitas, dan akurasi masing-masing sebesar 82,31%, 93,67%, dan 90,33%,. Sedangkan pada ROI bawah didapatkan rata-rata sensitivitas, spesifisitas, dan akurasi masing-masing sebesar 95,01%, 96,50%, 95,29% dan 2,59 detik.
==================================================================================================================
Dental panoramic radiograph is the x-ray image of two-dimensional (2-D) of the dental record an entire mouth, including the teeth, upper jaw, lower jaw and surrounding tissues and structures in the image. In dental panoramic radiograph contains a lot of information that can be identified one of them through the trabecular bone structure.
This research proposes the segmentation of trabecular bone area on dental panoramic radiographs based segment profile characteristics and classification methods Extreme Learning Machine. Input from these methods is dental panoramic radiographs. Selection of region of interest (ROI) was performed on the lower jawbone on trabecular bone area in which there are teeth and cortical bone. The ROI is divided into two where the upper ROI containing the teeth and lower ROI contains cortical bone. After that, the result of cutting ROI done preprocessing using mean and median filter to the upper ROI and motion blur for lower ROI. The image that has separated each extracted each pixel into 4 features consisting of image intensity, 2D Gaussian filter with two different sigma, and Log Gabor filter for upper-ROI. For lower ROI, used 5 feature extraction namely the intensity image, 2D Gaussian filter with two different sigma, phase congruency, and Laplacian of Gaussian. Then use some of the sample pixels as training data to create models of Extreme Learning Machine. The output of this classifier is a segmentation of trabecular bone area.
On top ROI, obtained an average sensitivity, specificity, and accuracy of respectively 82.31%, 93.67%, and 90.33%. While at lower ROI obtained an average sensitivity, specificity, and accuracy of respectively 95.01%, 96.50%, 95.29% and 2.59 seconds.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | dental panoramic radiograph, Extreme Learning Machine, Trabecular bone, segmentasi, segmentation |
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science. EDP R Medicine > RK Dentistry |
| Divisions: | Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
| Depositing User: | - RIZQI OKTA EKOPUTRIS |
| Date Deposited: | 05 Apr 2017 07:10 |
| Last Modified: | 06 Mar 2019 04:48 |
| URI: | http://repository.its.ac.id/id/eprint/3147 |
Actions (login required)
 |
View Item |