Med-LIFE: A System for Medical Imagery Exploration Joshua New Erion Hasanbelliu Knowledge Systems Lab JN 5/29/2016 Introduction • • • • What is Med-LIFE? What is image fusion? How do I teach the computer? How can I view the results? Knowledge Systems Lab JN 5/29/2016 What is Med-LIFE? • Med-LIFE is an application currently under development to use computer processing techniques to reduce medical personnel workload • GUI designed with Qt • Image Processing with C (and VTK library) Knowledge Systems Lab JN 5/29/2016 What is Med-LIFE? • Consists of three processes (LIFE): – Learning of image attributes by the computer using SFAM – Image Fusion of many image modalities into one color image – Exploration of learning and fusion results Knowledge Systems Lab JN 5/29/2016 What is Image Fusion? • Allows the combination of multiple image modalities into one colored image with no information loss • Reduces workload by eliminating the number of images a radiologist must analyze • Images used from “The Whole Brain Atlas” – http://www.med.harvard.edu/AANLIB/home.html Knowledge Systems Lab JN 5/29/2016 What is image fusion? • Technique similar to primate vision Knowledge Systems Lab JN 5/29/2016 Image Fusion Example PD GAD Color Fuse Result T2 JN 5/29/2016 SPECT Knowledge Systems Lab Image Fusion Example How do I teach the computer? • SFAM – Simplified Fuzzy ARTMAP • SFAM is a computer-based system capable of online, incremental learning • Two “vectors” are sent to this system for learning: – Input feature vector tells what data is available from which to learn – Supervisory signal tells whether that vector is an example or counterexample Knowledge Systems Lab JN 5/29/2016 How do I teach the computer? • Left-click to define examples (green) • Right-click to define counterexamples (red) Main Window Zoom Window Knowledge Systems Lab JN 5/29/2016 How do I teach the computer? • Supervisory signal from red/green marks • Feature vector from slice pixel values for original, single, and double opponent images Knowledge Systems Lab JN 5/29/2016 Learning Results Main Window Results Zoom Window Results Knowledge Systems Lab JN 5/29/2016 Learning Results Main Window Results JN 5/29/2016 T2 Knowledge Systems Lab How can I view results? • Display a plethora of information • Skull generated for patient from PD modality for contextual slice navigation • Explore tab provides several functions: – Original images – Fusion results imbedded within 3D, patientgenerated skull – Learning results Knowledge Systems Lab JN 5/29/2016 Demo Presentation • Erion will now demo the Med-LIFE system Knowledge Systems Lab JN 5/29/2016 Conclusion • Med-LIFE offers reduced workload to physicians who scan multiple images – Image processing and fusion reduces the number of images to be analyzed – Learning system allows the computer to perform prescreening or background analysis – Exploration allows immersion within the data for surgery planning Knowledge Systems Lab JN 5/29/2016