AI-Driven Classification of Tumor Operability using 3D-CT Scans
We developed a deep learning model which classifies the operability of tumors in 3D-CT scans using patient information and 3D-CT scans. The model is able to classify the operability of pancreas tumors in 3D-CT scans with an accuracy of 87%.
Highlights
- Novel deep learning model for tumor operability classification
- Classification of pancreas tumors in 3D-CT scans
- Preprocessing and postprocessing of 3D-CT scans
- Deep research into the field of medical imaging and deep learning
Facts
- Client: University of Basel
- Location: Zurich, Switzerland
- Year: 2024
- Use Case: Medical imaging and deep learning
This project, conducted in collaboration with the University of Basel, focuses on segmenting pancreatic tumors and related anatomical structures from 3D CT scans. It includes a full pipeline for preprocessing, training, validation, testing, and visualizing segmentation results. Key features include structured data handling, class balancing, and specialized evaluation metrics for tumor regions. The system is designed to support both development and analysis of medical image segmentation outcomes, with visual and quantitative outputs to aid performance monitoring.
