Sano Showcases SimuScope at WACV 2025 –  Advancing Synthetic Endoscopic Data Generation

Sano Showcases SimuScope at WACV 2025 – Advancing Synthetic Endoscopic Data Generation

Polish Innovation on the Global Computer Vision Stage Researchers from the Sano Centre for Computational Medicine, Joanna Kaleta and Sabina Martyniak, represented their team at the prestigious IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025. The event took place in Tucson, Arizona, and gathered leading experts in AI and image analysis. SimuScope: Advancing […]

Polish Innovation on the Global Computer Vision Stage

Researchers from the Sano Centre for Computational Medicine, Joanna Kaleta and Sabina Martyniak, represented their team at the prestigious IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025. The event took place in Tucson, Arizona, and gathered leading experts in AI and image analysis.

SimuScope: Advancing Medical AI with Realistic Synthetic Data

The team presented their poster titled “SimuScope: Realistic Endoscopic Synthetic Dataset Generation through Surgical Simulation and Diffusion Models”, which highlights a new method for producing synthetic medical imaging data. The research was conducted by Sabina Martyniak, Joanna Kaleta, Michał Naskręt, Szymon Płotka, and Przemysław Korzeniowski (the head of Sano’s Medical Imaging and Robotics group) and Diego Dall’Alba.

SimuScope addresses a pressing challenge in medical AI: the limited availability of high-quality, annotated endoscopic data. By combining surgical simulations with advanced diffusion models, the system creates realistic datasets that can be used for training and validating machine learning algorithms.

Open Access to Data, Code, and Methods

Whether you’re a scientist, developer, or simply curious about the future of computational medicine, SimuScope offers a comprehensive set of resources to explore:

  • A peer-reviewed paper detailing the methodology and findings,
  • Open-source code available for adaptation and testing,
  • A realistic synthetic endoscopic dataset ready for experiments and benchmarking.

Explore, test, and build on top of this work – it’s all publicly available: