The thesis is about segmenting neuronal growth-cones in phase-contrast time-lapse recordings. In this repository you'll find the Python- and R-code, which I used to process and evaluate my data.
To segment neuronal growth-cones one can use classic methods (such as Thresholding) or you can follow the trend of using artificial intelligence i.e. neural networks. I tested some classic methods including the power-law transformation and the logarithmic transformation.
See these links and paper for further information:
Links
Paper
Furthermore I tested the pix2pix-model a (generating adverserial network) GAN proposed by (Isola et al. 2016 ).
Please also visit the repository containing the source code I used to train and test the model:
[Torsten Bullmann]