Hand Shape Recognition using Higher Order Local Autocorrelation Features in Log Polar Coordinate Space

Abstract

The friendly communication can be more promoted between the human and computer if the function of gesture recognition is implemented to the computer system as the input interface along with the keyboards and mice. We propose a mouse-like function for estimating hand shape from input images with a monocular camera, with which a computer user feels no restraint or awkwardness. Our system involves conversion of sequential images from Cartesian coordinates to log-polar coordinates. Temporal and spatial subtractions and color information are used to extract the hand region. The origin of log-polar coordinates is chosen as the center of the acquired image, but once the hand has been extracted, the estimated centroid position of the hand region in the next frame, obtained from the current hand position and speed, is used as the origin to convert. Recognition of the hand shape is carried out by multiple regression analysis using higher order local autocorrelation features of log-polar coordinate space. Mouse-like functions are realized according to the hand shape and motion trajectory. Compared to conventional Cartesian coordinates, conversion to log-polar coordinates enables us to reduce image date and computation time, remove the variability by the scaling, and improve antinoise characteristics.

Real time search for similar hand images from database for robotic hand control

“…Except for equivalent pattems due to parallel shifting, the high-order local autocorrelational patterns can be expressed in 25 different kinds [4]. However, since the pattems M1 to M5 become smaller in value compared with other patterns, we squared the number of pixels at the reference point for M1, and further multiplied it with the number of pixels at the reference point for M2 to M25.…”

Realtime estimation of human hand posture for robot hand control

“…Except for equivalent patterns due to parallel shifting, the highorder local autocorrelational patterns can be expressed in twenty five different kinds [4]. However, since the patterns MI to M5 become smaller in value compared with other patterns, we squared the number of pixels at the reference point for MI, and further multiplied it with the number of pixels at the reference point for M2 to M25.…”