Kinematic Motion Estimation

Kinematic motion estimation is set of techniques for trying to track in real-time three dimensions, the movement of parts of an object, usually a user. It is necessary to estimate the position change rather than track it directly because artefacts are relatively common in real-time video capture systems, and because even if the view is perfect, tracking something like a finger moving through three dimensional space, is going to be subject to a degree of error owing to the time lapse, as well as telling the system nothing about how the rest of the body has moved to enable that finger to be there at that time.

In general, all methods work by analysing the difference between frames, and then using machine vision algorithms to first reconstruct what the body parts actually are, along with non-relevant data in the visual field. Once that is done, the identifiable parts are tracked, whilst any that are not identifiable / separable from the background / blurred / hidden are reconstructed, using a variety of possible methods, primarily (but not always) revolving around matching the user to a boned model of their joints or presumed joints.

The position of identifiable body parts and how that position has changed between frames, is then enough data to use to estimate quite reliably the position of body parts that are not currently recognisable or visible based on how the person's body would have to deform at the joints in order to reach that pose, as well as the type of motion they had been undergoing in the frames leading up to this point.

It is a complex process, but a very necessary one, as it is, in essence, replicating the human capability to read body language, mannerisms, and most importantly control gestures by the interface system.

All forms of gesture interface use some degree of kinematic motion estimation, and the process is particularly vital in motion capture systems, where the end goal is to translate the user's precise body language into a virtual environment to bring life to an avatar form.

See Also: Optoelectronic, Kinematics, Posturography, Mocap, Gait Analysis, Limb Kinematics, Kinematic Chain, Kinematic Pair, Forward Kinematics, Gait Analysis, Individual Bone Constraints, Inverse Kinematics, Bone Inverse Kinematics, Bone Weighting

Below, we offer a selection of links from our resource databases which may match this term.

Bone Inverse Kinematics

Kinematic Motion Estimation

Limb Kinematics

Results by page [1]   [2]   [3]   


Resource Type not Available

Cheapening the Cost of Motion Sensors
Motion sensors are starting to creep into a whole plethora of applications. They are the linch pins of haptics, of 3D pointers, of stress based sensor networks and locomotive VR interfaces. Yet, there's a problem. Small, discrete motion sensors, tiny enough to be built into larger devices the size say, of a Wii-remote or a 6 ounce HMD, are extremely difficult and expensive to produce.

Animals in Motion

Eadweard Muybridge's pioneering work on the motion of all manner of quadrupeds. Horses, goats, cats, gnus, eagles, gazelles, sloths, camels, many others shown walking, running, flying, leaping, almost a complete range of natural motions duplicated for the benefit of animators and skeletal model making.

The Male and Female Figure in Motion: 60 Classic Photographic Sequences

The paperback 'lightweight version of 'The Human Figure in Motion', this book as might be extrapolated from the title, contains 60 photographic sequences of the more common everyday activities of both men and women. Most are nude, allowing the musculature and natural drooping from one frame to another to be clearly seen and analysed.

Horses and Other Animals in Motion: 45 Classic Photographic Sequences

Horses and other Animals in Motion is a collection of, as the title says, 45 sets of photographs of horses hauling, walking, trotting, etc., plus sequences of donkeys, an ox, pig, dog, cat, deer and other animals capture details of anatomy and movement. These images, were taken by the definitive expert in the field, Eadweard Muybridge.

The Human Figure in Motion

After photographer Eadweard Muybridge created his revolutionary photographs of animals in motion in the late 1890s, he turned his attention to the study of the human form, by taking detailed photographs in rapid succession step by step as the human body underwent all manner of daily activities. These photographs have served for over a century, as the most highly acclaimed reference point for animators.

Muybridge's Animals in Motion CD-ROM and Book

This book and CD is essentially the electronic format version of Eadweard Muybridge?s ?Animals in Motion?. It contains electronic format versions of 167 black-and-white photographic sequences captures the movements of 34 different animals as they run, fly, leap, and perform other characteristic actions. Includes 10 bonus Flash animations plus 15 photographic sequences that are ready to be animated.

Results by page