The Leap Motion system recognizes and tracks hands and fingers. The device operates in an intimate proximity with high precision and tracking frame rate and reports discrete positions and motion.
The Leap Motion controller uses optical sensors and infrared light. The sensors are directed along the y-axis – upward when the controller is in its standard operating position – and have a field of view of about 150 degrees. The effective range of the Leap Motion Controller extends from approximately 25 to 600 millimeters above the device (1 inch to 2 feet).
Detection and tracking work best when the controller has a clear, high-contrast view of an object’s silhouette. The Leap Motion software combines its sensor data with an internal model of the human hand to help cope with challenging tracking conditions.
The Leap Motion system employs a right-handed Cartesian coordinate system. The origin is centered at the top of the Leap Motion Controller. The x- and z-axes lie in the horizontal plane, with the x-axis running parallel to the long edge of the device. The y-axis is vertical, with positive values increasing upwards (in contrast to the downward orientation of most computer graphics coordinate systems). The z-axis has positive values increasing toward the user.
The Leap Motion API measures physical quantities with the following units:
|Time:||microseconds (unless otherwise noted)|
As the Leap Motion controller tracks hands and fingers in its field of view, it provides updates as a set – or frame – of data. Each Frame object representing a frame contains any tracked hands, detailing their properties at a single moment in time. The Frame object is essentially the root of the Leap Motion data model.
The hand model provides information about the identity, position, and other characteristics of a detected hand, the arm to which the hand is attached, and lists of the fingers associated with the hand.
Hands are represented by the Hand class.
The Leap Motion software uses an internal model of a human hand to provide predictive tracking even when parts of a hand are not visible. The hand model always provides positions for five fingers, although tracking is optimal when the silhouette of a hand and all its fingers are clearly visible. The software uses the visible parts of the hand, its internal model, and past observations to calculate the most likely positions of the parts that are not currently visible. Note that subtle movements of fingers tucked against the hand or shielded from the Leap Motion sensors are typically not detectable. A Hand.confidence() rating indicates how well the observed data fits the internal model.
More than two hands can appear in the hand list for a frame if more than one person’s hands or other hand-like objects are in view. However, we recommend keeping at most two hands in the Leap Motion Controller’s field of view for optimal motion tracking quality.
An Arm is a bone-like object that provides the orientation, length, width, and end points of an arm. When the elbow is not in view, the Leap Motion controller estimates its position based on past observations as well as typical human proportion.
The Leap Motion controller provides information about each finger on a hand. If all or part of a finger is not visible, the finger characteristics are estimated based on recent observations and the anatomical model of the hand. Fingers are identified by type name, i.e. thumb, index, middle, ring, and pinky.
Fingers are represented by the Finger class.
A Finger object provides a Bone object describing the position and orientation of each anatomical finger bone. All fingers contain four bones ordered from base to tip.
The bones are identified as:
- Metacarpal – the bone inside the hand connecting the finger to the wrist (except the thumb)
- Proximal Phalanx – the bone at the base of the finger, connected to the palm
- Intermediate Phalanx – the middle bone of the finger, between the tip and the base
- Distal Phalanx – the terminal bone at the end of the finger
This model for the thumb does not quite match the standard anatomical naming system. A real thumb has one less bone than the other fingers. However, for ease of programming, the Leap Motion thumb model includes a zero-length metacarpal bone so that the thumb has the same number of bones at the same indexes as the other fingers. As a result the thumb’s anatomical metacarpal bone is labeled as a proximal phalanx and the anatomical proximal phalanx is labeled as the intermediate phalanx in the Leap Motion finger bone model.
(Original diagram by Marianna Villareal.)
Along with the computed tracking data, you can get the raw sensor images from the Leap Motion cameras.
The image data contains the measured IR brightness values and the calibration data required to correct for the complex lens distortion. You can use the sensor images for augmented reality applications, especially when the Leap Motion hardware is mounted to a VR headset.
For more information, see Camera Images.