Deploy Deep Learning Models on Raspberry Pi

Abstract

This chapter provides guidelines for deploying deep learning models on the Raspberry Pi, emphasizing real‐time artificial intelligence (AI) and computer vision applications. It demonstrates practical applications of various computer vision tasks on this popular embedded platform. This chapter explores face detection and recognition, demonstrating face detection using the histogram of oriented gradients (HOG) method and face recognition through transfer learning techniques. Real‐time object detection is covered with a focus on deploying the you only look once, version 3 (YOLOv3) model for accurate and fast object detection. The chapter also discusses real‐time image classification using PyTorch's quantized models, optimized for Raspberry Pi's limited resources. Additionally, real‐time segmentation with K‐means clustering is introduced to segment images effectively. This chapter has exercise problems that guide readers in applying some of the techniques presented and provide a hands‐on approach to deploying trained models on Raspberry Pi.