Search

A segmentation method that partly mimics the human cognitive-behavioral process of recovering motor-temporal information from the image of handwritten words is presented. Its main feature is that it can manipulate information regarding the curvature function of the handwritten word contour. The process makes it possible to detect lost or ambiguous parts of the image and interpret them to recover a part of the original temporal information.

This article reviews the current status of the movement primitives problem, especially the question of their description and identification, and points out some challenges which remain unsolved by the approaches frequently adopted to study human movements. We use the framework of the kinematic theory proposed by Plamondon (1995 and 1998) as an example of a hybrid model that allows a precise and flexible quantitative description of human movements. It is shown that the theory can be used to encompass the various types of rapid movement encountered in the field. The principal aim of this article is to highlight the fact that the notion of movement primitives should not necessarily be confined to movements with a single velocity peak, as is still often assumed in many models. The model allows, for example, a simple description of a movement primitive, which might contain up to two direction reversals.