Road Detection Example

CN24 ships with a very simple pre-trained model for road detection. For training, we used the KITTI-Vision Road dataset. You can use this model for fine-tuning as well, but in this short demonstration we will use it to classify images!

First, build CN24 with the JPEG option. Copy the files in the example folder into your build directory. Now, simply run the classifyImage tool:

$ ./classifyImage kitti_um_road.set kitti.net kitti_pretrained.Tensor sample1.jpg output1.jpg

If everything worked correctly, your output should look like this: