Dataset and libraries used
In this chapter, we will be using TensorFlow (v1.4.0) and OpenCV as our main library for detection. We show results on custom images. However, any colored image can be used as input for various models. Wherever required, there are links to pre-trained model files in the sections.