ABSTRACT
This chapter presents an image classification case study problem using deep learning. It uses an open-source road sign dataset that has four different classes: traffic light, stop sign, crosswalk, and speed limit. A deep learning model is trained on this dataset to classify these items correctly from images. The chapter covers some fundamental aspects of deep learning training in a practical scenario, such as setting default configurations (batch size, learning rate, or the number of training epochs, etc.) and defining the dataset class and its attributes to structure and handle the dataset effectively. The chapter also provides guidelines for setting up the model for multi-GPU and single-GPU workstations. The trained classification model is also tested using outside images for inference in the context of real-world applications.
