meshography workshop

Leveraging the meshing of frog gastrocnemius muscle from digital photography using artificial neural networks.

This data collection supports the finding in our paper submitted to the INTERNATIONAL JOURNAL OF NUMERICAL METHODS IN BIOMEDICAL ENGINEERING", on Oct.20,2018.

There are two folders (left and right) containing 15 images each, processed for edge detection of the left and right object contours.

There are also some Matlab scripts and functions, for data preparation and training of a Bayesian backpropagation neural network.

The script (kod2.m) is associated with the polynomial fitting procedure to the contours, and it utilized the function (fiterror.m) in the process.

The file (polinomtablo4th.mat) contains the output of the fitting procedure, and is used by the neural networking script (plotneur4th.m).

Finally, the unit cylinder mesh are processed through the network, to produce the 1767 node, 1440 hexahedral-type element FEAP mesh (wIplant2), of the muscle shown in (mesh_view2.png).

The codes are based in the MATLAB R15b, image processing and neural network toolboxes.