Chapter 5: Indoor Scene Classification with TensorFlow Lite for Microcontrollers and the Arduino Nano

Computer vision is what made convolutional neural networks hugely popular. Without this deep learning algorithm, tasks such as object recognition, scene understanding, and pose estimation would be really challenging. Nowadays, many modern camera applications are powered by machine learning (ML), and we just need to take the smartphone to see them in action. Computer vision also finds space in microcontrollers, although with limitations given the reduced onboard memory.

In this chapter, we will see the benefit of adding sight to our tiny devices by recognizing indoor environments with the OV7670 camera module in conjunction with the Arduino Nano 33 BLE Sense board.

In the first part, we will learn how to acquire images from the OV7670 camera module. We will then focus on the model design, applying transfer learning with the Keras API to recognize kitchens and bathrooms. Finally...