It will train a binary svm classifier to detect car objects in images. Training images will be of size 40*100 and test image can be of any size.
Sliding window operation is used to search car objects in test images and a bounding box is drawn around each detection.