$ cd ~
$ python3.9 -m venv <environment-name>
$ source <environment-name>/bin/activate
$ pip install -r <path-to-repo>/requirements.txt
- Images : Raw Image Data
- Labels : Labels as RGB Images
- Labels_int : Labels transfered as Integer Classes
- Test : One Test Driving Session, used to create an Video to visualize the Segmentation
- David_Description.pdf : Paper of the used Dataset
- data_visualization : Exemplaric Images and Histogram of the Occurrence of the Labeled Classes
- evaluation : Plots to illustrate the different Results of each Model
transform_data.ipynb
: Transfer RGB Labels of the Dataset into Integer Classesalexnet_transfer_learning.ipynb
: Train Unet with Alexnet as Encoderresnet18_transfer_learning.ipynb
: Train Unet with Resnet18 as Encodershufflenet_transfer_learning.ipynb
: Train Unet with Shufflenet as Encodermodel_architecture.py
: Decoder for the Resnet18evaluate_models.ipynb
: Notebook to compare the Results of the different Modelscreate_video.ipynb
: Notebook to visualize the Segmentation Results as Video- models : Already Trained Models to reload and use them without retraining
Johannes Mock
Simeon Grossmann
The used Dataset is provided by:
The Decoder used for the Resnet18 Model is taken from:
GNU GENERAL PUBLIC LICENSE Version 3