Details

Real-Time Sequence Data Analysis in 3D Slicer Using Deep Neural Networks

Year: 2025

Term: Winter

Student Name: Zeynep Kocacenk

Supervisor: Matthew Holden

Abstract: This project embeds an integrated system for real-time sequence data analysis into 3D Slicer—a widely used, free, open-source platform for envisioning healthcare intervention data. Time series analysis in this domain is still done offline, and immediate feedback for surgeons is extremely limited. This work integrates the 3D Slicer Sequences module with machine learning libraries like PyTorch or Keras to make real-time DNN analyses possible that will enable a range of functions: surgical coaching, automated visualization, and decision support. The project's objective is an interface for continuous data processing that can provide immediate, data-driven insights during interventions. In order to achieve this, the DNNs will be trained on data, such that real scenarios are emulated by the model. It is expected that, at the end of this process, a workable real-time analysis system in 3D Slicer may be realized, and health professionals can enhance both procedural accuracy and patient outcomes with immediate, actionable feedback. The innovation seeks to allow for better and more informed decisions right in the operating environment.