3D image reconstruction where the third dimension is time or depth from exotic light sensors or different light treatments (phase contrast) is addressing new and exciting imaging challenges which require non-invasive, eye-safe and/or low cost solutions. However there can be issues with data acquisition in terms of detection, sparsity and background noise. Also, large scale inverse problems generate significant computational challenges for real-time high resolution reconstruction. Here we look at different mathematical imaging methods for a range of physical and medical applications: cell tracking and modelling, imaging/depth prediction in low/obscured visibility conditions. Methods/Tools to be discussed will range from deep learning, ray tracing, circular Hough transform, variational segmentation and tracking methods. The presentations will include illustrations on realistic data sets. The results advance the state of the art in both reconstruction rate and quality.

Real-time Depth Prediction with Sensor Fusion.

Roderick Murray-Smith (University of Glasgow)

Mathematical Imaging Methods for Mitosis Analysis in Live-Cell Phase Contrast Microscopy.

Joana Grah (Graz University of Technology)

Imaging Behind Walls.

Piergiorgio Caramazza (University of Glasgow)

Photoacoustic Tomography Reconstruction from Sparse Data using Ray Tracing.

Francesc Rullan (University College London)

Organizers:

Catherine Higham (University of Glasgow)

Roderick Murray-Smith (University of Glasgow)

Keywords:

computed tomography, deep learning, image reconstruction, image segmentation, inverse problems, machine learning, partial differential equation models