For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Extraction of Personal Features from On-Line Handwriting Information in Context-Independent Characters
Yasushi YAMAZAKI Naohisa KOMATSU
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences
Publication Date: 2000/10/25
Print ISSN: 0916-8508
Type of Manuscript: Special Section PAPER (Special Section on Information Theory and Its Applications)
Category: Identity Verification
feature extraction, writer verification, handwriting, identity verification,
Full Text: PDF(663.4KB)>>
We propose an extraction method of personal features based on on-line handwriting information. Most recent research has been focused on signature verification, especially in the field of on-line writer verification. However, signature verification has a serious problem in that it will accept forged handwriting. To solve this problem, we have introduced an on-line writer verification method which uses ordinary characters. In this method, any handwritten characters (i.e., ordinary characters) are accepted as a text in the verification process, and the text used in the verification process can be different from that in the enrollment process. However, in the proposed method, personal features are extracted only from the shape of strokes, and it is still uncertain how efficient other on-line information, such as writing pressure or pen inclination, is for extracting personal features. Therefore, we propose an extraction method of personal features based on on-line handwriting information, including writing-pressure and pen-inclination information. In the proposed method, handwriting information is described by a set of three-dimensional curves, and personal features are described by a set of Fourier descriptors for the three-dimensional curves. We also discuss the reliability of the proposed method with some simulation results using handwritten data. From these simulation results, it is clear that the proposed method effectively extracts personal features from ordinary characters.