In this repo you find all scripts and relevant files we created for our project "CNN-based regression of drop impact parameters" which was part of the Data Driven Engineering 2 lecture held by Cihan Ates. The project data was given by Maximilian Dreisbach. Thank you Cihan and Max for your support!
-
get_parameters.py is used to extract all features we created from the render data given by Maximilian Dreisbach. It was submitted to the HPC by using a bash script, the extracted parameter data was saved in a json file.
-
get_angles_from_pictures.ipynb validates our angle calculation (which was performed via trimesh in get_parameters.py) using a grafical approach.
-
parameter_calculation.ipynb is the long version of get_parameters.py including explanations and validations of the introduced parameters.
-
cnn_project.ipynb is the notebook where we did the CNN-based regression of our drop impact parameters, using input data with and without glarepoints. In this context we saved the best model trained on input data with glarepoints as well as the best model trained on input data without glarepoints to reuse it for later visualization.
-
cnn_project_second_meeting.ipynb is a former version of the cnn_project.ipynb but was saved seperately, as it contains single output prediction models, whereas the final project concentrates on multi output prediction only.
-
visualization.ipynb visualizes what happens in the CNN layers of the models created in our cnn_project.ipynb script.
##Contributions
This project was created in the course of the DDE2 project and was coded in collaboration with two other course attendees.