DFRobot SEN0670 Gravity AI Pose/Gesture Recognition Sensor
DFRobot SEN0670 Gravity AI Pose and Gesture Recognition Sensor is an offline AI vision module that enables custom pose and hand-gesture recognition without requiring data collection or AI model training. Users can teach and store up to eight custom gestures and eight full-body poses using Windows® configuration software. The sensor detects 17 body keypoints at approximately 10FPS and 21 hand keypoints at approximately 5FPS for responsive human-machine interaction. Learned actions are stored locally and communicated through I2C or UART interfaces to Arduino, ESP32, and micro:bit platforms.
Designed for embedded applications, the DFRobot SEN0670 operates without internet connectivity and delivers reliable local inference for gesture- and pose-based control systems. The compact 37mm × 37mm module features low-power operation, a dedicated wake-up pin, USB Type-C® configuration, and support for Gravity connectors and standard pin headers. The sensor is ideal for interactive art installations, smart toys, gesture-controlled lighting, music systems, educational projects, and pose-controlled displays. Flexible custom learning capabilities allow developers to create personalized interfaces and interactive experiences with minimal development effort.
Features
- Custom static pose learning - store up to 8x custom full-body poses locally without training an AI model
- Custom static hand gesture learning - store up to 8x custom single-hand gestures for personalized interaction
- 17x body keypoints - human pose mode runs at approximately 10FPS with a recommended distance of around 2m
- 21x hand keypoints - hand gesture mode runs at approximately 5FPS with a recommended distance of around 0.5m
- Switchable recognition modes - gesture and pose recognition operate as separate modes and can be switched through the Windows software or supported commands
- Offline local inference - recognition runs on the sensor without cloud AI or an internet connection
- Standalone deployment after learning - learned actions are stored locally and can be used without keeping the PC connected
- Low-power mode - approximately 40mA operating current at 3.3V and approximately 37μA sleep current at 3.3V, with WAKEUP control
- Maker-friendly interfaces - I2C, UART, Gravity PH2.0-4P, 2.54mm header, and USB Type-C® configuration
- Development support - Arduino IDE and MakeCode support, compatible with Arduino, ESP32, and micro:bit through the appropriate connection hardware
- Windows configuration software - supports Windows 10/11 for live preview, custom learning, mode switching, and firmware updates
- CE-EMC and RoHS certified
- For ideal recognition results
- Use even indoor lighting and avoid strong backlights
- Keep the hand or the full body unobstructed
- Keep the background relatively simple when possible
- Use clearly differentiated custom gestures or poses
- For hand interaction, keep the hand around 0.5m from the sensor
- For full-body pose recognition, keep the full body visible at around 2m
- Designed for static gesture and pose recognition, dynamic motion trajectories are not supported
Applications
- Full-body pose-controlled displays and projection
- Interactive art installations
- Gesture-controlled lighting and music projects
- Smart toys and desktop interactive devices
Specifications
- Recognition parameters
- 21x hand keypoints
- 17x human skeleton keypoints
- 13x fixed gestures
- Up to 8x custom gesture capacity, deletable and re-learnable
- Up to 8x custom pose capacity, deletable and re-learnable
- 0.1m to 1m gesture recognition range under adequate lighting (0.5m recommended)
- 2m to 4m pose recognition range under adequate lighting (2m recommended)
- ~5FPS gesture recognition frame rate
- ~10FPS skeleton recognition frame rate
- Hand recognition
- Single-hand only
- Custom gestures limited to one hand at a time, even with multiple hands present
- Pose recognition
- Multi-person supported
- Custom poses limited to one person at a time, even with multiple people present
- Recognition type - static actions only, dynamic actions not supported
- 640×480 camera resolution (VGA)
- 89° diagonal field of view (DFOV)
- 3.01mm camera effective focal length (EFL)
- Interface parameters
- I2C and UART communication
- PH2.0-4P 2.54mm pin header connector
- USB Type-C receptacle debug/configuration port
- 0x3A I2C address
- 921,600bps UART default baud rate
- Basic parameters
- 3.3V to 5V supply voltage
- 3.3V logic level
- 40mA typical operating current at 3.3V
- 37μA typical sleep current at 3.3V (low-power mode required)
- WAKEUP pin
- Physical dimensions
- 37mm × 37mm PCB size
- 30mm mounting hole spacing
- ø3.2mm mounting hole diameter
Kit Contents
- 1x Gravity AI pose and gesture recognition sensor
- 1x Gravity 4-pin sensor cable
- 1x 10-pin, 2.54mm black single-row pin header
Videos
Gesture Controls
Pinout
Dimensions
