Abstract: A method is provided for reconstructing data generated on a digitizer tablet based on the dynamics of the data generation. The digitizer tablet samples the position of the writing pen, continuously transmitting data to a computer in the form of x and y coordinates plus an indicator of whether or not the pen is touching the surface of the tablet. After preprocessing, the data is reconstructed based on the dynamics of the data's creation. The values of the dynamic reconstruction are created by accumulating increments that are some function of the original data. To enhance the dynamics a nonlinear function is required.

Abstract: A method is provided for reconstructing data generated on a digitizer tablet based on the dynamics of the data generation. The digitizer tablet samples the position of the writing pen, continuously transmitting data to a computer in the form of x and y coordinates plus an indicator of whether or not the pen is touching the surface of the tablet. After preprocessing, the data is reconstructed based on the dynamics of the data's creation. The values of the dynamic reconstruction are created by accumulating increments that are some function of the original data. To enhance the dynamics a nonlinear function is required.

Abstract: Processing of handwritten patterns, for example, inputting of complex characters, such as Kanji, into a word processor involves a balance between an algorithmic part and a disambiguation part. Application of routines uniformly during the algorithmic part deliberately results in confusion sets of characters which are subsequently differentiated during the disambiguation part. Balance between the two parts permits a flexibility to accommodate differences in writing styles, while application of routines customized to the established character patterns permits the unique solution necessary for accurate recognition of a character.

Abstract: A method and system for recognizing complex patterns, such as Chinese characters (or Kanji), identifies patterns by stroke recognition using detailed information about stroke shape and specifically curvature characteristics. Written strokes are categorized into different specific stroke types and are assigned stroke identity labels by comparison of distance metrics generated with reference to stroke templates, each template comprising information specifying position of a point and permissible deviation for each such point along a stroke. Deviation values may differ from point to point and from stroke type to stroke type. The data on the written stroke may be rotated or displaced relative to each stroke template to enhance the likelihood of stroke recognition. Limited information about stroke position relative to other strokes may be preserved and employed to resolve ambiguities among accumulations of stroke identity labels found to ambiguously indicate more than one pattern or character.

Abstract: A method for recognizing complex handwritten character set employs spatial position recognition in which characters of a first set are registered in a first region and characters of a second set are registered in a second region. The second region may be included within the first region. The invention is particularly useful for distinguishing Kanji from non-Kanji character sets in a data input system.