Comparison of two vehicle detection methods based on the oriented gradients histogram and the deep neural network

In this article, we present two approaches to mobile object recognition in a real-time video scene. The first method is based on deep learning to detect moving objects. The second method is based on extracting the HOG characteristics associated with each object. The results obtained by these two methods are interesting in the detection of moving objects. For both approaches, the classification phase is done by the Alexnet neuron network.