relation: https://digilib.uin-suka.ac.id/id/eprint/66480/
title: PENINGKATAN RESOLUSI CITRA FOTOAKUSTIK MELALUI METODE ENHANCED SUPER-RESOLUTION GENERATIVE ADVERSARIAL NETWORKS (ESRGAN)
creator: Handi Putra Utama, NIM.: 20106020032
subject: 530 Fisika
description: Photoacoustic imaging with the raster scanning method requires a long acquisition time and low image resolution results. Conventional interpolation methods commonly used in digital image processing, including neighborhood, bilinear, and bicubic, aren’t optimal enough because they still leave blur and noise effects. This research aims to enhance the resolution of photoacoustic images of dental samples through the ESRGAN method with Skyrim and Deoldify models. The process is carried out using an interface program that has been created. Thus, comparisons between images from two ESRGAN models and conventional interpolation methods could be evaluated. Neighborhood, bilinear, and bicubic interpolation works enhance the resolution of photoacoustic image resolution, which initially had the widest pixel density by FWHM (Full Width at Half Maximum) of 3,76 pixels achieved to 3,64, 3,55, and 3,49 pixels, as well as signal dominance against noise measured by SNR (Signal-to-Noise Ratio) of 2,17 dB to 3,68, 3,87, and 4,40 dB. Other side, ESRGAN can provide better resolution results where the Skyrim and Deoldify models have FWHM values of 3,02 and 2,99 pixels also SNR of 5,54 and 5,71 dB.
date: 2024-07-11
type: Thesis
type: NonPeerReviewed
format: text
language: id
identifier: https://digilib.uin-suka.ac.id/id/eprint/66480/1/20106020032_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf
format: text
language: id
identifier: https://digilib.uin-suka.ac.id/id/eprint/66480/2/20106020032_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf
identifier:   Handi Putra Utama, NIM.: 20106020032  (2024) PENINGKATAN RESOLUSI CITRA FOTOAKUSTIK MELALUI METODE ENHANCED SUPER-RESOLUTION GENERATIVE ADVERSARIAL NETWORKS (ESRGAN).  Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.