Alphanumeric handprint recognition
AssigneeScan-Optics, Inc.
InventorsDavid H. Stone, Leon A. Pintsov, Donald P. D'Amato
Life of this patent
Estimating the term…
Abstract
The recognition of patterns is accomplished through boundary tracing and subsequent storage of encoded range testable data commensurate with the occurrence, interrelationship and orientation of geometric features on the pattern boundaries. The encoded data is compared with generalized prototypes which define the geometric shape of all probable permutations of all possible patterns to be recognized, a match between the encoded data and a prototype constituting identification of the pattern.
Claims13
An identification process comprising the steps of:
- storing information commensurate with at least a portion of the boundary of an object to be identified;
- analyzing the stored information to ascertain the location of points on the object boundary which meet preselected curvature conditions, at least some of said ascertained points being points of peak curvature which lie within regions of preselected length and curvature on the object boundary;
- further analyzing the stored information to determine the geometric shape of the object boundary regions adjacent to the ascertained points;
- identifying each of the analyzed boundary regions as a feature of a preselected set of features on the basis of geometric shape;
- additionally analyzing the stored information to determine the position relative to the object as a whole of each identified feature;
- also analyzing the stored information to determine the direction of each feature relative to a predetermined orientation;
- preparing a range testable encoded description of the said object boundary portion in the entirety;
- preparing a range testable encoded description at the identified features on the said boundary portion and their interrelationship;
- generating a series of generalized descriptions which define the geometric shape of all probable permutations of the objects to be recognized, said generalized descriptions including both desired and undesired features and being in the said range testable code;
- comparing the encoded descriptions with generalized descriptions of the series of generalized descriptions;
- and
- noting the identify of the object when the encoded descriptions match a generalized description.
The method of claim 6 wherein the step of creating a binary two dimensional representation comprises:
- producing an image of a portion of a document, the image including at least a first pattern to be identified;
- and
- digitizing and storing the image.
A process for the identification of a pattern imprinted upon a document comprising the steps of:
- creating and storing a binary two dimensional representation of an image of a pattern to be identified;
- completely tracing the exterior boundary of the stored image representation in a first direction, the trace thus encompassing boundary regions having curvature;
- identifying regions on the traced boundary which have a curvature which satisfies preselected criteria including magnitude and type, regions of convex curvature being identified by detecting points of peak curvature therein;
- classifying each identified region on the basis of geometric shape;
- determining the orientation of each classified region relative to a predetermined coordinate system;
- preparing and storing a range testable encoded description of the boundary of the pattern to be identified as a function of the shapes of the identified regions, their orientation and their relative position on the traced boundary;
- generated and storing a compact range testable description of each topologically unique prototype of each pattern to be identified, said prototype description including information commensurate with regions of both desired and undesired shape;
- and
- serially comparing the range testable encoded description of the boundary of the pattern to be identified with the prototype descriptions and noting the identify of the pattern when the descriptions match.
Dependencies are inferred from the claim text.