Handwriting recognition and comparison, though previously and frequently implemented, has rarely supported input from multiple users without distinct prior identification of the active user providing input. Adaptive algorithms can be incorporated with existing methods of handwriting identification in order to compare arbitrary samples with pre-recorded database entries of font file-like handwriting samples which have been identified with their writers. Such a program would allow for reliable computational comparison usable in environments such as court situations and banking.

Our project's goal is to implement such an algorithm in order to match handwriting input within a database of stored handwriting samples.

We intend to model analog handwriting analysis techniques, using a time parameter to identify how letters are written as a means of comparison in addition to shape comparison with individual letters. The general methods used by handwriting analysts to identify who wrote a letter based on the motions involved in its writing would first be used to identify which database file to use, and then the letters would be identified by shape and motions.

We have written C++ objects to manage bitmaps (with member functions to binarize and thin letter images to single pixels so that they may be better identified), database files (capable of storing user identification data and standardized letter image positioning, in order to arbitrary determine the location and size of individual letters), and letters in our program and have begun to delve into the important steps of implementing the input and identification routines. We may require a touch input device in order to continue with programming our projects, since a time parameter is virtually impossible to implement without realitme input.

Considering our limited progres thus far and the magnitude of the steps ahead, it may be unreasonable for us to expect project completion by the deadline. However, assuming we do complete the project, we expect to be able to identify several fairly distinct handwriting samples from one another with ease.