Data Science Programming All-in-One For Dummies

Chapter 2

Performing Complex Classifications

IN THIS CHAPTER

Discovering publicly available image datasets

Augmenting images in various ways

Delving into traffic sign recognition

Deep neural network solutions have been highly successful in the image-recognition field. The great part of this technology’s success, especially in AI applications, comes from three key characteristics: the availability of suitable data to train and test image networks; the application of deep neural networks to different problems thanks to transfer learning; and increasing sophistication of the technology, which allows it to answer complex questions about image content.

In this chapter, you delve into the topic of object classification and detection challenges to discover their contribution to the foundation of the present deep learning renaissance. Competitions, such as those based on the ImageNet dataset, provide the right data to train reusable networks for different purposes (thanks to transfer learning, as discussed in Book 4, Chapter 3). Competitions serve another purpose as well, which is to push researchers to find new and smarter solutions for increasing the a neural network’s capability to ...