ESP32-CAM Face Recognition using Edge Impulse

This project implements a localized machine learning system that captures visual data, processes it, and provides real-time feedback. At the core of the circuit is the ESP32-CAM module, which integrates an ESP32 microcontroller with an OV2640 camera sensor. The camera captures raw image frames from the environment. Because the ESP32 features built-in Wi-Fi and Bluetooth along with a dual-core processor, it can run a lightweight, quantized machine learning model directly on the edge—trained via Edge Impulse—without needing an external cloud server. When the system detects a specific face within the camera's field of view, it classifies the image data according to the trained neural network. Concurrently, the system communicates with a 0.91-inch OLED display, typically using the I2C communication protocol, to visually output system status or classification results directly to the user. To program and supply code to the ESP32-CAM (which lacks an onboard USB port), a USB to UART converter module is connected via female-to-female jumper wires to the board's serial pins during the flashing process. Once deployed, the system operates autonomously, leveraging machine learning inference to recognize faces in real time while displaying feedback on the compact screen.