Zigzag persistent homology for processing neuronal images
Authors:
Gadea Mata,
Miguel Morales,
Ana Romero, and
Julio Rubio
Abstract:
We apply the ideas of zigzag persistence to determine the objects of interest in stacks of neuronal images, locating and marking different dendrites. In particular, this allows us to recognize some 3D properties of the objects, distinguishing dendrites that cross, but not intersect, in the ambient space. The algorithms are implemented in a Fiji/ImageJ plugin, usable on two different kinds of images.
Reference:
Gadea Mata, Miguel Morales, Ana Romero, Julio Rubio, Zigzag persistent homology for processing neuronal images, Pattern Recognition Letters, Volume 62, 1 September 2015, Pages 55-60, ISSN 0167-8655, https://doi.org/10.1016/j.patrec.2015.05.010.
Images transfected with Actin-GFP
These type of images depict the neuronal structure of rat hippocampal neurons in culture (inmunocytochemistry). They are transfected with an Actin-GFP expression vector - following Morales et al. (2000).
You can download this kind of image here: images for the experiment 1.
Images stained with DiI
These images (inmunohistochemistry) depict CA1 pyramidal hippocampal neurons from a transgenic hippocampal mice stained with DiI - as described in Enríquez-Barreto et al. (2014).
You can download this kind of image here: images for the experiment 2.
Validation process
Report explaining the validation process of the plugin NeuronZigzagJ: NeuronZigzagJvalidation.pdf.