This project is undertaken by a PhD student at the Responsible Artificial Intelligence Lab and it aims to create efficient deep learning algorithms for processing gigapixel images, which are extremely large and present significant computational challenges. The research seeks to improve the accuracy and computational efficiency of deep learning techniques, focusing on applications in computational pathology. The aim is to enhance the analysis of gigapixel images, such as whole slide images (WSIs), for medical research and clinical diagnosis.
Deep learning has transformed image processing, but gigapixel images common in fields like pathology are too large for conventional deep learning methods. These images are critical in medical research as they contain intricate details necessary for accurate diagnosis. However, they are extremely large and prone to artefacts (e.g., tissue folds, staining errors) that degrade image quality and diagnostic accuracy. This research investigates the development of more general and efficient models to address the computational bottlenecks of high-resolution image processing.
The research focuses on the need for efficient processing of high-resolution images in medical diagnosis. Its findings will contribute to the progress of computational pathology and other areas that rely on high-resolution imaging, such as remote sensing and surveillance.

Albert Dede (PhD)