Diagnosing Medical Images with Agentic AI

This master's thesis focuses on developing an AI agent for the automatic diagnosis of Optical Coherence Tomography (OCT) images. The project involves creating comprehensive knowledge bases, designing an effective workflow, and implementing an intelligent system. It aims to enhance diagnostic accuracy and efficiency in medical imaging, leveraging advanced artificial intelligence techniques for potential improvements in patient care outcomes.

Informationen
- Masterarbeit für 1 Person
- 30% Theorie, 70% Umsetzung
Voraussetzungen
- Basiswissen in KI
- Interesse an medizinischen Prolemstellungen
- Python von Vorteil
Objective
We are seeking an innovative and motivated master student to develop an AI agent designed to automatically evaluate and diagnose Optical Coherence Tomography (OCT) images. This project aims to push the boundaries of medical imaging diagnostics by leveraging advances in artificial intelligence to streamline and enhance the diagnostic process in ophthalmology and other fields.
Project
The primary objective of this master’s thesis is to create an intelligent system capable of accurately diagnosing medical images, specifically OCT scans, by integrating sophisticated AI techniques. The project will involve several key steps:
1. Knowledge Base Development
  - Generate one or more comprehensive knowledge bases that serve as the foundation for the AI agent’s decision-making capabilities. This will include:
    - Harnessing image datasets from renowned ophthalmology experts.
    - Incorporating Natural Language Processing (NLP) data regarding OCT.
    - Utilizing resources from other resources, e.g., the National Center for Biotechnology Information (NCBI) and/or PubMed Central.
    - Integrating patient metadata
2. Workflow Definition
  - Design and define an effective workflow that outlines the process through which the AI agent will operate. This workflow will be crucial in ensuring the agent’s efficiency, accuracy, and reliability in various diagnostic scenarios.
3. AI Agent Implementation
  - Develop and implement the AI agent, focusing on its ability to autonomously assess and diagnose OCT images. This will involve selecting appropriate algorithms, training models, and fine-tuning parameters to optimize diagnostic accuracy.
4. Agent Evaluation:
  - Conduct a thorough evaluation of the AI agent’s performance. This will include testing its diagnostic abilities against established benchmarks and iteratively improving the system based on evaluation results. Emphasis will be placed on achieving a balance between sensitivity and specificity in diagnostic outcomes.
The latter stages of the project, from workflow design to evaluation, are expected to be iterative. Repeated testing and refinement will be essential for honing the AI agent’s capabilities and ensuring its alignment with clinical needs and standards.
Who Should Apply
This project is ideally suited for students with a background in AI, machine learning, computer vision, or related fields. Familiarity with medical imaging and its applications is a plus. Candidates should possess a keen interest in artificial intelligence and its potential to revolutionize medical diagnostics.
Benefits
Participants will gain hands-on experience at the intersection of AI and healthcare, working with cutting-edge technology and datasets. This project offers an opportunity to make substantive contributions to medical AI and potentially improve patient outcomes through innovative diagnostic tools.
Application Process
Interested candidates are invited to submit their resumes along with a short text detailing their interest in the project and relevant experience. Please send your applications to thesis@scs.ch or use the form below.