Marwan Abdellah

To help segmenting the neuronal skeletons into axons, dendrites and spines, we should provide complete statistical analysis data and apply those statistics to the skeletonization algorithm at the last stage.

This task tests the limits of the skeletonization algorithm implemented in Ultraliser. __Process__ * We must be able to detect if the resulting skeleton is multi-partitioned or has a single...

An important milestone to detect the spines from the skeleton and then find the corresponding geometry in the input mesh (or volume). __Process__ * The geometry is re-voxelized with better...

To improve the performance of the vascular morphologies it is better to implement a reader that is compatiable with the vmv5 format that is similar to the vmv format but...

To improve the performance of the volume-based functions, implement a fast WOBC versions of the indexing functions.

Please add to the documentation the structure of the header files that are loaded for raw data.

Implement edge collapse-based mesh simplification.

2

To reduce the mesh size while preserving the topology based on a cost function, we should implement the edge-collape algorithm in Ultraliser. An existing implementation is found here https://github.com/andandandand/progressive-mesh-reduction-with-edge-collapse.

For neurons with small details and low volume occupancy, we need sparse voxel octree implementation to be able to overcome the memory limitation required for establishing huge volume grids.

For verification and visual illustration purposes, we should implement morphology exporters in H5 and SWC formats.

In certain cases, the soma is not given in the morphology file, and also we cannot reconstruct a valid configuration with which we can build any soma. We therefore should...