Datasets and libraries
We will be continuing the use of OpenCV and NumPy for image processing. For deep learning, we will use Keras with the TensorFlow backend. For segmentation, we will be using the Pascal VOC
dataset. This has annotations for object detection, as well as segmentation. For tracking, we will use the MOT16
dataset, which consists of an annotated sequence of images from video. We will mention how to use the code in the sections where it is used.