Gravity Offline Edge AI Gesture/Face Detection Sensor

The DFRobot gravity offline edge AI gesture and face detection sensor is designed for intuitive gesture control and multi-person presence detection. Equipped with an integrated camera and onboard AI processor, the sensor can recognize five predefined hand gestures and detect up to ten faces or upper-body presences simultaneously within a 3 meter range. By combining computer vision and real-time edge AI computing, it delivers responsive and private interaction without cloud dependency.

Featuring dual communication interfaces (I²C and UART) and wide voltage support (3.3 V to 5 V), this module is fully compatible with Arduino, Raspberry Pi, ESP32, and graphical platforms like MakeCode® and Mind+. It is ideally suited for smart home interfaces, touchless control systems, public interaction devices, and privacy-focused IoT solutions.

All recognition tasks are handled by the onboard AI processor where no data is stored or transmitted to the cloud. This completely offline operation guarantees data privacy, making it ideal for sensitive applications in healthcare, public spaces, or homes. Despite its powerful capabilities, the offline AI gesture sensor has a low operating current of just 100 mA, making it perfect for power-sensitive and standalone projects.

The offline AI gesture and face detection sensor accurately recognizes five common hand gestures from 0.5 meters to 3 meters away. Each recognized gesture triggers a corresponding colored RGB light, providing instant visual feedback for seamless interaction. This makes it perfect for contactless control of smart appliances, interactive installations, and assistive devices.

The five recognized gestures are: Good (thumbs-up, blue light), OK (OK gesture, green light), Stop (open palm, red light), Victory (V sign, yellow light), and Call me (call me gesture, purple light).

Beyond gestures, the offline AI gesture and face detection sensor excels at detecting human presence by identifying the head-and-shoulders area. It can track up to ten people simultaneously and outputs the total count in real-time. For more advanced applications, it provides the position coordinates (X/Y) and confidence score of the most clearly detected face. This enables sophisticated functions like occupancy-based climate control, people-counting for retail analytics, or a fan that automatically tracks a person's position.

With an 85° diagonal field of view and a 3 meter detection range, the long-range AI gesture and face detection sensor ensures consistent and reliable performance in a variety of environments. Whether mounted on a wall, integrated into a device, or placed on a desk, its wide coverage is ideal for everything from small rooms to larger, multi-user spaces like classrooms or lobbies.