Index

Hierarchy-Aware Deep Learning for Tironian Notes Recognition

Deep Learning-Based Classification and Explainability of Cytomegalovirus Encephalitis in Longitudinal MRI Data

This Master Project focuses on developing and evaluating advanced deep learning methodologies for the automated detection and classification of Cytomegalovirus (CMV)-induced encephalitis using clinical Magnetic Resonance Imaging (MRI) data.

Motivation and Goal

CMV encephalitis is a challenging condition to diagnose, and advanced, non-invasive computational methods are required to assist clinical decision-making. The primary goal is to leverage the temporal and multi-modal information within longitudinal MR scans to classify the presence or stage of inflammation (encephalitis).

Data and Scope

The project utilizes a unique, high-quality, pre-selected longitudinal dataset comprising MRI scans from approximately 300 patients, with an average of six scan time points and multiple MR sequences available per visit.

Key Tasks and Research Questions

Literature Review: Conduct a targeted literature review of similar projects focused on MR classification, specifically those dealing with longitudinal data (e.g., the BraTS Challenge: Predicting the Tumor Response During Therapy) and methods for medical image classification and prediction.
Model Development: Adapt and implement state-of-the-art deep learning architectures (e.g., 3D Convolutional Neural Networks, Recurrent Neural Networks, or hybrid models) suitable for processing longitudinal and multi-sequence volumetric data.
Explainability (XAI): A critical component of the project is the integration of Explainable AI techniques (e.g., Grad-CAM, Saliency Mapping, or LRP). The student will implement and evaluate these methods to highlight which anatomical regions or temporal patterns contribute most significantly to the model’s classification decision, thereby increasing clinical trust and interpretability.
Evaluation: The core research question is whether robust classification performance can be achieved on the available data, considering potential constraints such as fewer advanced sequences or less acute disease stages compared to published reference literature. Performance will be measured using metrics like AUC, sensitivity, and specificity.

Development of a Local LLM Agent System for Clinical Expert Support and Automation in MRI Planning for Radiation Therapy

RAG-Enhanced Low-Cost Vision-Language Models for Diabetic Retinopathy Classification and Automated Reporting

Diabetic Retinopathy (DR) affects over 160 million people globally, projected to reach 180 million by 2030, despite 90% of related blindness being preventable through early detection [1]. Current AI models achieve strong classification performance but lack interpretable clinical reports, limiting their adoption in low-resource settings. Although Vision-Language Models (VLMs) offer unified diagnosis and report generation, fundus captioning significantly underperforms compared with other imaging modalities [2,3], and state-of-the-art VLMs remain computationally expensive. Although Retrieval-Augmented Generation (RAG) has improved medical imaging accuracy [4], no prior study has integrated DR severity grading, lesion-aware reporting, and evidence retrieval within a low-cost, clinically deployable VLM.

Generalization of Self-Supervised Vision Models in Image Retrieval

This project investigates the generalization capabilities of prominent self-supervised vision models, such as DINOv2, CLIP, and MoCo, when applied to image retrieval tasks across diverse visual domains.

Overview

Self-supervised learning (SSL) models are increasingly important for creating robust visual features. However, their performance often degrades when transferring from standard natural image datasets (like ImageNet) to more specialized domains.

Our study systematically benchmarks these models on three distinct domains:

1. Natural Images (Baseline)

2. Scene-Centric Images (e.g., Places365)

3. Artistic Images (e.g., ArtPlaces)

We compare the models’ out-of-the-box generalization and analyze the impact of finetuning on their domain adaptation, providing crucial insights into the stability and robustness of the learned representations for practical retrieval applications.