20.500.12556/RUL-161364 Uporaba superresolucije za izboljšanje kakovosti nizkoresolucijskih slik Use of super-resolution for enhancing the quality of low-resolution images Diplomska naloga obravnava uporabo superresolucije za izboljšanje kakovosti nizkoresolucijskih slik. Superresolucija je postopek, pri katerem se iz slike nizke ločljivosti ustvarijo slike z višjo ločljivostjo. V nalogi je bil obravnavan problem nizke kakovosti naravnih klasičnih slik, kar lahko vpliva na različne aplikacije. Glavni cilj naloge je bil izboljšati kakovost teh slik s pomočjo naprednih modelov superresolucije, kot sta Real-ESRGAN in ResShift. Metodologija dela je vključevala pregled teoretičnih osnov superresolucije, razvoj in implementacijo modelov Real-ESRGAN in ResShift ter njuno testiranje na nizkoresolucijskih slikah. Uporabljene so bile različne metrike za merjenje uspešnosti superresolucije, kot so PSNR, SSIM, BRISQUE in NIQE, s katerimi je mogoče kvantitativno oceniti kakovost izboljšanja resolucije vhodnih slik. Rezultati naloge so pokazali, da lahko uporabljeni modeli znatno izboljšajo kakovost nizkoresolucijskih slik. Bolj natančno: model ResShift na splošno bolje deluje pri metrikah PSNR in SSIM v primerjavi z modelom Real-ESRGAN, kar nakazuje boljšo kakovost rekonstrukcije. The thesis deals with the use of super-resolution to improve the quality of lowresolution images. Super-resolution represents an image processing task that aims to transform a low-resolution image into a high-resolution image. In the thesis, we addressed the problem of low quality of input images, which can negatively impact the performance of various downstream computer vision tasks and applications. The main goal of the thesis was to improve the quality of these images using advanced super-resolution techniques such as Real-ESRGAN and ResShift. The methodology of the work included an overview of the theoretical foundations of super-resolution, development and implementation of Real-ESRGAN and ResShift models, and their testing on low-resolution images. Various quality metrics such as PSNR, SSIM, BRISQUE, and NIQE were used to objectively evaluate the capabilities of the tested superresolution techniques. The results of our experiments showed that the used models can significantly improve the quality of low-resolution images. More specifically, the ResShift model generally performs better on the PSNR and SSIM metrics compared to the Real-ESRGAN model, that it can ensure better reconstruction quality. superresolucija Real-ESRGAN ResShift kakovost slike super-resolution Real-ESRGAN ResShift image quality true false false Slovenski jezik Angleški jezik Diplomsko delo/naloga 2024-09-10 12:05:01 2024-09-10 12:05:12 2024-09-12 03:25:47 0000-00-00 00:00:00 2024 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 62520 207132419 Blazheski_David_-_Uporaba_superresolucije_za_izboljsanje_kakovosti_nizkoresolucijskih_slik.pdf Blazheski_David_-_Uporaba_superresolucije_za_izboljsanje_kakovosti_nizkoresolucijskih_slik.pdf 1 292935989AC5052F3BC69333B39EC755 28a61f6b71f06610996d069d7ec010cfbcf9d8dcb99a06e62e852dad43855a5e 2547f701-6f5c-11ef-b232-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=190281 Fakulteta za elektrotehniko 0 0 0