Udacity Self-Driving Car Engineer Nanodegree Program – Term 3, Project 2: Semantic Segmentation

This repository contains the code written to complete the second project in term 3 of the Udacity Self-Driving Car Engineer Nanodegree Program. This project uses semantic segmentation to label the pixels of a road in images using a Fully Convolutional Network (FCN).

The load_vgg() method downloads a pre-trained VGG16 model and extracts the input, keep probability, layer 3, layer 4, and layer 7 (main.py, lines 21-46). Using these extracted layers the layer() method creates the rest of the Fully Convolutional Network (FCN):

Every convolutional layer and transposed convolutional layer uses an initializer that generates a truncated normal distribution with a standard deviation of 0.01 and a L2 kernel regularizer with a scale of 0.001.

After the network is defined the optimize() method creates the optimizer and the cross entropy loss using an Adam Optimizer (main.py, lines 91-110).

Finally, the network is trained using the train_nn() method (main.py, lines 114-144) with a keep probability of 0.5 and a learning rate of 0.00001. During training log messages describe the behaviour of the loss value.

The network has been trained using the Kitti Road dataset (289 images) from here. The network training has been run for 6, 12, 24, and 48 epochs using a batch size of 8. The following table

6 epochs
12 epochs
24 epochs
48 epochs

6 epochs
12 epochs
24 epochs
48 epochs

6 epochs
12 epochs
24 epochs
48 epochs

6 epochs
12 epochs
24 epochs
48 epochs

6 epochs
12 epochs
24 epochs
48 epochs

6 epochs
12 epochs
24 epochs
48 epochs

The description below is Udacity's original README from the seed project.

In this project, you'll label the pixels of a road in images using a Fully Convolutional Network (FCN).

main.py will check to make sure you are using GPU - if you don't have a GPU on your system, you can use AWS or another cloud computing platform.

Make sure you have the following is installed:

• Python 3
• TensorFlow
• NumPy
• SciPy

Download the Kitti Road dataset from here. Extract the dataset in the data folder. This will create the folder data_road with all the training a test images.

Implement the code in the main.py module indicated by the "TODO" comments. The comments indicated with "OPTIONAL" tag are not required to complete.

Run the following command to run the project:

python main.py

Note If running this in Jupyter Notebook system messages, such as those regarding test status, may appear in the terminal rather than the notebook.

1. Ensure you've passed all the unit tests.
2. Ensure you pass all points on the rubric.
3. Submit the following in a zip file:
   • helper.py
   • main.py
   • project_tests.py
   • Newest inference images from runs folder (all images from the most recent run)

• The link for the frozen VGG16 model is hardcoded into helper.py. The model can be found here.
• The model is not vanilla VGG16, but a fully convolutional version, which already contains the 1x1 convolutions to replace the fully connected layers. Please see this forum post for more information. A summary of additional points, follow.
• The original FCN-8s was trained in stages. The authors later uploaded a version that was trained all at once to their GitHub repo. The version in the GitHub repo has one important difference: The outputs of pooling layers 3 and 4 are scaled before they are fed into the 1x1 convolutions. As a result, some students have found that the model learns much better with the scaling layers included. The model may not converge substantially faster, but may reach a higher IoU and accuracy.
• When adding l2-regularization, setting a regularizer in the arguments of the tf.layers is not enough. Regularization loss terms must be manually added to your loss function. otherwise regularization is not implemented.

If you are unfamiliar with GitHub , Udacity has a brief GitHub tutorial to get you started. Udacity also provides a more detailed free course on git and GitHub.

To learn about REAMDE files and Markdown, Udacity provides a free course on READMEs, as well.

GitHub also provides a tutorial about creating Markdown files.