Syuhada, Fahmi Syuhada (2020) Segmentasi Multi Proyeksi Pada Citra Cone Beam Computed Tomography Gigi Menggunakan Metode Level Set. Masters thesis, Institut Teknologi Sepuluh November.
Preview |
Text
05111850010037-Master_Thesis.pdf Download (3MB) | Preview |
Abstract
Segmentasi citra Cone Beam Computed Tomography (CBCT) gigi
merupakan suatu penelitian yang menarik untuk dilakukan. Metode segmentasi berbasis pelacakan kontur seperti metode level set merupakan metode yang banyak digunakan saat ini. Umumnya penelitian yang sudah dilakukan yaitu mengembangkan metode level set untuk dapat mendeteksi keseluruhan objek gigi dengan baik. Salah satu hal yang menjadi perhatian dari penelitian sebelumnya yaitu proses segmentasi yang hanya menggunakan data citra CBCT gigi dengan tipe proyeksi aksial. Dengan menggunakan proyeksi ini, metode dirancang agar seoptimal mungkin untuk mendapatkan bentuk keseluruhan. Namun, susunan citra irisan pada proyeksi aksial yang merepresentasikan bagian ujung objek gigi mengalami penurunan kontras sehingga membuat objek gigi cenderung tidak dapat dibedakan denan latar ataupun unsur lain. Oleh sebab itu dibutuhkan integrasi atau kombinasi proyeksi Sagital dan Koronal yang beberapa citra irisan penyusunnya
menampilkan bentuk gigi secara penuh. Tesis ini mengusulkan metode segmentasi multi-proyeksi dengan mengkombinasikan hasil segmentasi level set pada citra dekomposisi CBCT gigi
untuk mendapatkan keseluruhan bagian objek gigi. Dekoposisi terhadap citra CBCT dilakukan untuk mendapatkan susunan citra irisan pada tiga proyeksi yaitu Aksial, Sagital, dan Koronal. Segmentasi kemudian dilakukan pada keseluruhan
citra irisan pada masing-masing proyeksi menggunakan metode level set. Hasil akhir dari metode usulan didapatkan dari kombinasi keseluruan citra irisan yang sudah tersegmentasi pada tiga proyeksi tersebut. Berdasarkan hasil uji coba, Metode segmentasi multi-proyeksi citra CBCT gigi menggunakan level set dengan mendekomposisi data citra CBCT berhasil
digunakan untuk mendapatkan keseluruhan bentuk objek gigi dengan akurasi 97.18%, sensitifitas 88.62 %, dan spesifisitas 97.61%. Kombinasi tiga proyeksi (segmentasi multiproyeksi) yaitu proyeksi Aksial dengan Sagital dan Koronal dapat memperbaiki hasil segmentasi proyeksi aksial yang kurang dalam menghasilkan keseluruhan bagian gigi dengan besar rata-rata peningkatan hasil dari 96.87%
menjadi 97.18%.
========================================================================
Dental Cone Beam Computed Tomography (CBCT) image segmentation is an interesting study to do. Contour tracking based segmentation methods such as the level set method are the most widely used methods today. Generally the research that has been done is to develop a level set method to be able to detect overall dental objects properly. One of the concerns of previous research is the segmentation process that only uses dental CBCT image data with axial projection types. Using this projection, the method is designed to be as optimal as possible to get the overall shape. However, the composition of the sliced image on the axial projection that represents the end of the tooth object has decreased contrast, making the dental object tends to be indistinguishable from the background or other elements.
Therefore, it requires integration or a combination of Sagital and Coronal projections in which some of the constituent image slices display the full shape of
the teeth. This reserch proposed a multi-projection segmentation method by combining the segmentation result using level set in decomposition of dental CBCT image to obtain the entire dental object. Decoposition of the CBCT image was carried out to obtain three projections namely Axial, Sagital, and Coronal. The segmentation is then performed on the entire sliced image in each projection using level set method. The final result is obtained from the combination of segmentation result on the three projections images. Based on the results of the experiment, the multi-projection segmentation method of dental CBCT images using level sets by decomposing the CBCT image
data was successfully used to obtain the overall shape of dental objects with an accuracy of 97.18%, specificity of 88.62%, and specificity of 97.61%. The combination of three projections (multiprojection segmentation) between Axial and
Sagonal and Coronal projections can improve the results of axial projection segmentation that is less in producing the whole tooth with an average increase in yield from 96.87% to 97.18%
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Segmentasi, Gigi, CBCT, Multi-Proyeksi, Level set, Dekomposisi, Proyeksi, Kombinasi |
| Subjects: | Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7888.3 Digital computers |
| Divisions: | Faculty of Information Technology > Informatics Engineering > 55101-(S2) Master Thesis |
| Depositing User: | Fahmi Syuhada |
| Date Deposited: | 16 Aug 2020 01:46 |
| Last Modified: | 05 Jun 2023 15:08 |
| URI: | http://repository.its.ac.id/id/eprint/78358 |
Actions (login required)
 |
View Item |