Akademik Veri Yönetim Sistemi
3rd International Conference on Artificial Intelligence, Computer, Data Sciences, and Applications, ACDSA 2026, Boracay Island, Filipinler, 5 - 07 Şubat 2026, (Tam Metin Bildiri)
In medical diagnostic applications, ultrasound (US) imaging is widely used owing to its portability, non-invasiveness, and real-time capability. Nevertheless, US B-mode images are often limited by relatively low resolution and speckle noise in the background. This study proposes a modified encoder-decoder deep learning approach aimed at enhancing the resolution of US B-mode images. The network was trained and evaluated on three US datasets, enabling robust assessment across varying acquisition conditions. The results of this study demonstrate that the proposed encoder-decoder architecture effectively improves image resolution while mitigating speckle, thereby potentially facilitating more accurate diagnostic interpretation. These findings underscore the potential clinical utility of deep learning-based super-resolution in ultrasound and lay the groundwork for future validation in broader patient populations and diverse imaging environments.