Image analysis – Pattern recognition – Unconstrained handwriting
Reexamination Certificate
1999-04-22
2001-02-06
Chang, Jon (Department: 2723)
Image analysis
Pattern recognition
Unconstrained handwriting
C382S219000
Reexamination Certificate
active
06185332
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to pattern and cipher recognition, particularly where a hand-written character has been digitized and is presented in the form of a pixel matrix.
Known character recognition systems achieve accuracy by either placing restrictions on the writer or by requiring substantial computer processing time and/or memory. Current character recognition techniques typically include normalizing and thinning an image corresponding to a character before determining image representation. This typically requires use of an extensive database for template matching necessitating a memory consuming dictionary to find possible matches between the image and words. The computer resources required reduces the efficiency of these techniques in computers having limited computational power. Some techniques for recognition of hand-written characters necessitates formation of the characters with a predetermined sequence of strokes, often having predetermined start and end points. Other techniques require that characters be formed with a single continuous stroke. The aforementioned constraints associated with prior art hand-written character recognition techniques requires users to create characters unnaturally. This often increases the time required for a user to become proficient with a particular character recognition system. The user must become accustomed to the constraints and learn to create characters in accordance with the same.
What is needed, therefore, is a character recognition technique that substantially reduces the constraints imposed upon a user when creating characters.
SUMMARY OF THE INVENTION
The present invention includes a method, a computer product and a system, each of which features a static recognition technique that facilitates accurate character recognition while substantially reducing the constraints placed on an end-user. To that end, a hand-written character, referred to herein as a cipher, is overlayed with a matrix having a plurality of cells with each cell having a plurality of pixels corresponding thereto. A subset of the pixels of a subgroup of the plurality of cells have a sub-portion of the cipher associated therewith, and it is determined whether the number of pixels associated with each portion of the cipher are equal. If not, cells having excess pixels compared to other cells in the matrix, have the excess pixels assigned to a different cell. This facilitates creating a digital word that describes the cipher that is then compared with information in a look-up table. The bits in the digital word correspond to a binary state of each of the cells in the matrix, i.e., whether a logical “0” corresponds to a cell or a logical “1”. An additional digital word is created that corresponds to the spatial position within the matrix where excess pixels are assigned. The information in the look-up table includes a plurality of known ciphers, each of which includes a corresponding digital representation. The digital information of each entry in the look-up table is compared with the digital word. A fuzzy logic algorithm determines the closest matches in the look-up table, and the cipher that corresponds to the closest match is identified. If more than one cipher in the look-up table is identified as a match, the digital representations corresponding thereto are processed and chosen as a function of the number of times the same cipher is identified as a match. The known cipher that is determined to match the digital word is provided, as needed. In other words, a visual representation of the known cipher may be presented on a display, or the known cipher may be stored in either a volatile of non-volatile memory.
These and other embodiments of the present invention, as well as its advantages and features, are described in more detail in conjunction with the text below and attached figures.
REFERENCES:
patent: 4155072 (1979-05-01), Kawa
patent: 4510617 (1985-04-01), Mori
patent: 4556985 (1985-12-01), Hongo
patent: 4712242 (1987-12-01), Rajasekaran et al.
patent: 5271068 (1993-12-01), Ueda et al.
patent: 5291560 (1994-03-01), Daugman
patent: 5319721 (1994-06-01), Chefalas et al.
patent: 5325447 (1994-06-01), Vogt, III
patent: 5438631 (1995-08-01), Dai
patent: 5479533 (1995-12-01), Tanaka
patent: 5515455 (1996-05-01), Govindaraju et al.
patent: 5802204 (1998-09-01), Basehore
patent: 5802207 (1998-09-01), Huang
patent: 5805727 (1998-09-01), Nakano
patent: 5805747 (1998-09-01), Bradford
patent: 5812698 (1998-09-01), Platt
patent: 5838820 (1998-11-01), Bergman
patent: 5841901 (1998-11-01), Arai et al.
Thompson Archie L.
Van Saun Richard A.
Allen Kenneth R.
Chang Jon
Faxtrieve, Inc.
Townsend and Townsend / and Crew LLP
LandOfFree
Character recognition technique does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Character recognition technique, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Character recognition technique will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2571551