Super Resolusi Pada Citra Mikroskopis Menggunakan Metode Enhanced Super Resolution Generative Adversarial Network (Esrgan)

Azizah, Ika Puteri Nur (2022) Super Resolusi Pada Citra Mikroskopis Menggunakan Metode Enhanced Super Resolution Generative Adversarial Network (Esrgan). Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06111840000022-UNDERGRADUATE_THESIS.pdf] Text
06111840000022-UNDERGRADUATE_THESIS.pdf
Restricted to Repository staff only

Download (4MB) | Request a copy

Abstract

Pencitraan mikroskop optik konvensional pada dasarnya dibatasi oleh produk space-bandwidth (SBP) dari sistem optiknya yang biasanya dalam bentuk megapiksel. Hal ini mengakibatkan adanya kontradiksi antara mencapai resolusi tinggi dan mempertahankan bidang pandang (FOV) yang besar. Padahal, citra mikroskopis resolusi tinggi dengan bidang pandang yang lebar saat ini dibutuhkan para ahli medis untuk melakukan penelitian. Salah satu contoh pengaplikasian citra mikroskopis pada bidang medis adalah untuk pemeriksaan hematologi berupa Hitung Darah Lengkap (Counting Blood Cell/CBC). Pentingnya CBC ini adalah untuk mengetahui kondisi kesehatan tubuh secara menyeluruh, mendiagnosis penyebab gangguan kesehatan atau penyakit, dan untuk memantau perkembangan penyakit bagi pasien yang sudah terdiagnosis mengalami gangguan kesehatan. Teknik untuk mendapatkan keseimbangan antara resolusi tinggi dan bidang pandang yang lebar pada citra mikroskopis diantaranya yaitu menggunakan teknik bernama Resolusi Super Gambar Tunggal (Single Image Super Resolution/SISR). Pada penelitian ini diimplementasikan teknik SISR pada citra mikroskopis sel darah menggunakan metode Enhanced Super Resolution Generative Adversarial Networks (ESRGAN). Hasil rata-rata tertinggi kinerja yang diukur menggunakan PSNR diperoleh saat menerapkan ESRGAN pada citra mikroskopis skala perbesaran 2 kali adalah 44,176, pada skala perbesaran 4 kali adalah 38,553, dan pada skala perbesaran 31,749 adalah 35,026.
==============================================================================================================================
Conventional optical microscope imaging is basically limited by the space-bandwidth (SBP) product of its optical system which is usually in the form of megapixels. This results in a contradiction between achieving high resolution and maintaining a large field of view (FOV). In fact, high-resolution microscopic images with a wide field of view are currently needed by medical experts to conduct research. One example of the application of microscopic images in the medical field is for hematological examinations in the form of a Complete Blood Count (Counting Blood Cell/CBC). The importance of this CBC is to know the overall health condition of the body, to diagnose the cause of health problems or diseases, and to monitor the progress of the disease for patients who have been diagnosed with health problems. Techniques to achieve a balance between high resolution and wide field of view in microscopic images include using a technique called Single Image Super Resolution (SISR). In this study, the SISR technique was implemented on microscopic images of blood cells using the Enhanced Super Resolution Generative Adversarial Networks (ESRGAN) method. The highest average result of performance measured using PSNR was obtained when applying ESRGAN to microscopic images at 2 times magnification scale was 44,176, at 4 times magnification scale was 38,553, and at 31,749 magnification scale was 35,026.

Item Type: Thesis (Other)
Additional Information: RSMa 006.42 Azi s-1 2022
Uncontrolled Keywords: CBC, Citra Mikroskopis, ESRGAN, SISR. CBC, ESRGAN, Microscopic Images, SISR.
Subjects: Q Science > QA Mathematics
Divisions: Faculty of Mathematics and Science > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User: Mr. Marsudiyana -
Date Deposited: 05 Jun 2026 02:01
Last Modified: 05 Jun 2026 02:01
URI: http://repository.its.ac.id/id/eprint/133588

Actions (login required)

View Item View Item