Probabilistic Generative Frameworks for Sampling 3D Complex Shapes and Images

Explore a cutting-edge statistical framework for simulating realistic 2D and 3D shapes in this insightful lecture. Delve into the challenges of developing computational tools for detecting global patterns in morphological variation using large-scale 3D surface scan databases. Learn about sub-image selection techniques and their importance in identifying physical features that best describe variations between classes of 3D objects. Understand the limitations of current shape simulation methods and discover how this new probabilistic generative approach addresses these issues. Examine two practical applications in computational biology: cellular imaging of neutrophils and mandibular molars from primate suborders. Gain valuable insights into the potential of this framework for advancing research in fields where access to real shape data is limited or expensive to collect.

Lorin Crawford 19/04/23: Probabilistic Generative Frameworks for Sampling 3D Complex Shapes & Images