Using deep learning models in image processing

Xiong, Z.

Xiong, Z. (2026)

Using deep learning models in image processing

Doctoral Thesis

Deep learning has revolutionized image processing, enabling remarkable advances in fields from autonomous driving to medical diagnostics. Yet the very success of these models, fueled by increasing complexity, has raised urgent questions about how they work, why they sometimes fail, and whether they can be trusted in critical applications. This thesis addresses these questions by offering both a coherent map of the rapidly evolving landscape of deep learning in computer vision and concrete contributions that advance model interpretability and reliability. The thesis is structured as a logical progression from a broad, systematic overview to focused investigations that together build a foundation for more transparent and robust vision systems.

Deep Learning

Convolutional Neural Network

Vision Transformers

Variational Bayes Inference

Conditional Probabilistic Model

Renal Pathology

Multi-label Lesion Classification