Peningkatan Kualitas Citra CT Scan Otak Menggunakan Metode DWT-CLAHE Untuk Deteksi Awal Stroke Iskemia

Hapsari, Andina Diya Ayu Paramita (2020) Peningkatan Kualitas Citra CT Scan Otak Menggunakan Metode DWT-CLAHE Untuk Deteksi Awal Stroke Iskemia. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
07311640000025_Undergraduate Thesis.pdf - Accepted Version
Restricted to Repository staff only

Download (4MB) | Request a copy

Abstract

Stroke adalah penyakit yang mempengaruhi pembuluh darah yang menyuplai darah ke otak. Stroke dapat menyebabkan disabilitas jangka panjang atau bahkan dapat menyebabkan kematian. Sekitar 85% dari kasus stroke merupakan stroke iskemia. Pengenalan gejala-gejala stroke secara cepat merupakan kunci dari penanganannya. Pencitraan memiliki peranan penting dalam diagnosis stroke. Computed Tomography (CT) adalah prosedur emergency pertama yang dilakukan. Dalam pencitraan CT, stroke iskemik terlihat sebagai daerah gelap (hypodense) dengan kontras relatif terhadap daerah sekelilingnya. Terdapat beberapa metode untuk meningkatkan kualitas kontras pada citra CT Scan, namun pada penelitian ini metode yang digunakan yaitu kombinasi antara Discrete Wavelet Transform (DWT) dengan Contrast Limited Adaptive Histogram Equalization (CLAHE). Parameter Entropi Diskrit, Strutctural Similarity (SSIM) dan Universal Image Quality Index (UIQI) digunakan sebagai analisa statistik kuantitatif untuk melihat peningkatan kualitas dari citra CT. Berdasarkan hasil penelitian, didapatkan hasil peningkatan citra yang paling baik yaitu dengan median filter kernel 3×3, penggunaan mother wavelet jenis Coiflet 2 serta dekomposisi level 1 pada DWT dan pemilihan clip limit sebesar 0,01 pada CLAHE. Sehingga didapatkan hasil rata-rata Entropi Diskrit sebesar 14,312, SSIM sebesar 0,987 dan UIQI sebesar 0,383. Dapat disimpulkan, metode DWT-CLAHE mampu menghasilkan intensitas cahaya yang tidak berlebihan namun memiliki kemampuan diferensiasi atau pembedaan kontras yang lebih baik antara jaringan lunak normal pada otak dan infark hypodense pada otak, serta terbukti dapat meningkatkan kualitas citra CT Scan asli yang memiliki rata-rata Entropi Diskrit sebesar 4,381. Untuk lebih meningkatkan hasil diagnosa, pada penelitian selanjutnya dapat dilakukan segmentasi daerah hypodense dengan memanfaatkan machine learning. ============================================================================================================= Stroke is a disease that affects the blood vessels that supply blood to the brain. Stroke can cause long-term disability or even cause death. About 85% of stroke cases are ischemic strokes. Rapid recognition of the symptoms of stroke is the key to handling it. Imaging has an important role in stroke diagnosis. Computed Tomography (CT) is the first emergency procedure performed. In CT imaging, ischemic stroke is seen as a dark area (hypodense) with contrast relative to the surrounding area. There are several methods to improve the contrast quality in CT Scan images, but in this study the method used is a combination of Discrete Wavelet Transform (DWT) with Contrast Limited Adaptive Histogram Equalization (CLAHE). The discrete Entropy, Structural Similarity (SSIM) and Universal Image Quality Index (UIQI) are used as a quantitative statistical analysis to observe the quality improvement of CT Scan images that have been improved compared to the original CT Scan images. Based on the result, the best image enhancement results were obtained with a median filter 3×3, the use of Coiflet 2 mother wavelets, level 1 decomposition on DWT and clip limit selection of 0.01 on CLAHE. So, we get an average result of Discrete Entropy of 14.312, SSIM of 0.987 and UIQI of 0.383. It can be concluded that the DWT-CLAHE method is able to produce intensity that is not excessive but has the ability to differentiate or distinguish better contrast between normal soft tissue of brain and hypodense infarction in the brain, and is proven to be able to improve the quality of the original CT Scan image that has an average Discrete Entropy of 4.381. To improve the diagnosis results in future study, the hypodense area segmentation can be done by utilizing machine learning.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Computed Tomography (CT), image enhancement, stroke iskemia, ischemic stroke
Subjects: T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques
Divisions: Faculty of Electrical Technology > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User: Andina Diya Ayu Paramita Hapsari
Date Deposited: 25 Aug 2020 03:05
Last Modified: 25 Aug 2020 03:05
URI: http://repository.its.ac.id/id/eprint/80579

Actions (login required)

View Item View Item