Challenges

Multivalued 3D+Time images acquired with cutting-edge microscopy and molecular engineering technologies, are complex for scientists to detect and visualize “on the fly” meaningful localized spatiotemporal events and processes in live cell or tissue imaging.

OBJECTIVE

We plan to develop machine-learning-based visualization, navigation, and interaction methods to investigate temporal series of multi-valued volumetric microscopy images.

TASKS AND ISSUES
  • Navigation and interaction based on machine learning in complex functional 3D+Time  data to visit sparse sets of localized intra-cellular events (fusion, fission, etc.) and cell processes (migration, division, etc.).
  • Quantification and analysis of sparse sets of molecular interactions and cell dynamics,  during the navigation to save memory and computational resources.
  • Visualization techniques of 3D motion vectors with appropriate non sub-resolved representations and discretization, as used to display 2D motion vectors.
APPLICATION DOMAINS

 

We investigate 3D+Time fluorescence Light Sheet Microscopy (LSM) – a new generation of modern microscopy – generates several thousands Gigabytes to several Terabytes / day !

  • 3D+Time LSM videos of growing organisms (animals,  plants) with low photoxicity.
  • 3D+Time Lattice LSM  for intracellular analysis of single cells or in small organisms.
GENERAL IMPACT
  • Academic framework combining machine learning, visualization, interaction, and image analysis methods.
  • Software for several communities, application fields and transfer to industry: 3D medical imaging, molecular and developmental biology, chemistry, fluid dynamics,…
  • Education: Visualization and navigation software for training biologist/microscopist users or for education purposes (License & Master levels).

Comments are closed.