Метод покращення якості зображень із низькою роздільною здатністю

Abstract

The paper considers methods for improving the quality of digital images. The focus is on classical image processing algorithms and modern deep learning methods. A combined approach is proposed, which integrates noise smoothing using classical techniques with image upscaling via the Real-ESRGAN neural network. The results of this work can be applied in computer graphics, medical imaging, video surveillance, and mobile applications.