How espressif/esp-who works
Project type and purpose
• This is an embedded image-processing platform (“ESP-WHO”) based on the Espressif ESP-IDF component model.
• It provides a collection of modular components (camera capture, ML inference, display drivers, peripheral drivers, application logic) and end-user examples (face detect, face recognition, QR-code, pedestrian detect) for various ESP32 series boards.
File structure
• Top-level folders:
• components/…: individual ESP-IDF components
• examples/…: complete projects tying components together
• tools/: build/CI helpers
• docs/, default_bin/: documentation and prebuilt firmware
• Inside components/, you see logical layers:
• who_peripherals: camera, LCD, SPI-flash, USB drivers
• who_frame_cap: frame capture pipeline
• who_detect, who_recognition, who_qrcode: core ML inference engines
• who_frame_lcd_disp: display pipeline
• who_app: application glue (“detect_app”, “qrcode_app”, “recognition_app”) + result handlers
• who_task: task scheduler / state machine
• Build files: each component has CMakeLists.txt and IDF metadata (idf_component.yml or Kconfig).
README
• Confirms ESP-WHO is built atop ESP-DL for asynchronous camera + model inference.
• Lists supported boards and features (camera, LCD, audio, buttons, etc.).
• Examples live under examples/… and use environment variable IDF_EXTRA_ACTIONS_PATH to pull in tools/.