US 4,093,941 A · Slope feature detection system

Abstract

A character feature detection system is provided for reliably reading non-touching hand printed and multi-font machine printed alphanumeric characters. An information field is scanned by a sensor array, and the sensor responses are digitized to form a binary character image. The image is thinned in one-cell layers as the image boundary and loops internal to the image are traced from boundary point to boundary point. A transition from one boundary point to another is defined by one of eight equally divergent vectors which are indicative of the local slope of a boundary between connected points. The sequence of vectors and their x-ray locations are recorded, and boundary slopes are indicated by a running average of four vectors representative of cell to cell transitions. Character features are detected from the sequence of boundary slopes to form a feature set including circumference counts, maximum and minimum image boundary limits, inflection points and loops. The image boundary is thinned to a one-cell stroke width, and character stops and nodes are located. Each stop arm is traced to identify valid stops, and the slope of each valid stop is determined to complete the feature set. The feature set is formatted and stored in a memory unit for character identification.

Claims14

Dependent claims

A method of detecting descriptive features of binary images of characters stored in a memory means formed from sensor cell responses generated by a sensor array scanning handprinted and multifont machine printed characters, which comprises:

(a) generating first signals representing the outer boundaries and inner loops of said characters by a vector control means in communication with said memory means;
(b) generating second signals which are vector representations of said boundaries by a trace control means receiving said first signals, each of said representations comprising one of a set of a plurality of equally divergent vectors;
(c) generating a first set of indicia representative of said characters representing averages of selected sets of said second signals by said trace control means;
(d) reducing each character to a one-cell stroke width by said vector control means and generating a second set of indicia representative of reduced ones of said characters from third signals produced by said vector control means tracing the outer boundary of said reduced ones.

1 more element

A method of detecting features of a character within a binary image formed by passing a sensor array through registration with either a handprinted or a machine printed character and stored in a memory means, which comprises:

(a) scanning said memory means to detect a first boundary point of said character by a scan control means;
(b) tracing from said first boundary point to successive boundary points of outer boundaries and inner loops of said character in stepwise moves by a vector control means which generates first signals according to said stepwise moves;
(c) accumulating a trace count of cell to cell moves taken during a boundary trace by a trace control means;
(d) comparing said trace count to an outer boundary threshold count and an inner loop threshold count by said trace control means;
(e) if said trace count is at least equal to said outer boundary threshold count or said inner loop threshold count, generating second signals forming a vector set representative of said cell to cell moves by said trace control means;
(f) generating boundary slope indicia as a function of said second signals by said trace control means;
(g) generating third signals comprising a slope feature set representative of detected characteristics in said boundary slope indicia by said trace control means;
(h) removing selected boundary points from said character in said memory means to form a reduced character having a one-cell stroke width by said vector control means;
(i) tracing said reduced character in said memory means and generating a spatial set of indicia representative of the occurrence of stops and nodes of said reduced character by said trace control means;
(j) tracing each of said stops of said reduced character and counting the number of cell to cell moves as a measure of stop length by said trace control means;
(k) removing those of said stops having a stop length less than a predetermined stop threshold count by said vector control means;
and
(l) generating fourth signals comprising a set of stop slope indicia representative of the direction of each of said stops in said memory having a stop length at least equal to said stop threshold count by said trace control means.

10 more elements

A method of detecting primary features of an unnormalized and multi-pieced binary character image comprising characters stored in a memory mosaic, which comprises:

(a) scanning said mosaic for a black cell with a white cell to its immediate left to locate an end of scan boundary point of each character satisfying a mark condition by a scan control means;
(b) generating a boundary trace of said character in stepwise moves from said end of scan boundary point by a vector control means;
(c) accumulating a count of cell to cell moves during said boundary trace in a counter, and comparing row and column coordinates of each boundary point encountered to detect maximum and minimum coordinates of said image in a trace control means;
(d) thinning said image in one-cell layers and marking boundary points encountered during said boundary trace which satisfy said mark condition by said vector control means;
(e) upon the occurrence of a boundary point satisfying an end of trace condition, scanning said mosaic by said scan control means from a last end of scan boundary point to detect a next end of scan boundary point as represented by an unmarked boundary point satisfying said mark condition;
(f) repeating steps (b) through (e) until no unmarked boundary points satisfying said mark condition are detected;
(g) if said count is at least equal to a character piece threshold count or an inner loop threshold count, generating first signals comprising a vector set representative of said stepwise moves;
(h) generating second signals forming slope indicia as averages of those members of said first signals representative of said cell to cell moves;
(i) generating third signals comprising slope feature indicia representative of said second signals;
(j) thinning said character to a one-cell stroke width, and tracing said character to detect stops and nodes by said vector control means;
(k) vertically ordering and tracing each of said stops by said trace control means;
(l) comparing a stop count of cell to cell moves encountered in tracing each of said stops with a threshold stop count by said vector control means;
and
(m) generating fourth signals comprising stop slope indicia representative of those of said stops having a stop count at least equal to said threshold stop count.

11 more elements

A system for detecting primary features of binary images of both handprinted and multifont machine printed aliphanumeric characters, which comprises:

(a) data control logic means responsive to electrical signals from independent sources to activate said system;
(b) plural image memory means in electrical communication with said data control logic means for storing said binary images;
(c) coordinate control logic means responsive to said data control logic means for controlling the loading of said binary images into said plural image memory means;
(d) scan control logic means responsive to said coordinate control logic means for scanning said binary images to detect a boundary point of a character satisfying a mark condition;
(e) trace enable logic means responsive to said scan control logic means for synchronizing boundary trace and feature data formatting operations;
(f) vector control logic means responsive to said trace enable logic means for tracing outer boundary and inner loops of said character and thinning said character to a one cell thickness, said vector control logic means including first means for generating vector representations of said outer boundary and inner loops, said trace enable logic means including second means responsive to said vector representations for averaging selected sets thereof to form vector averages;
(g) logic detector means in electrical communication with said vector control logic means for detecting the presence of character pieces and loops in said characters;
and
(h) feature memory means responsive to said scan control logic means for storing said primary features including said vector averages for character recognition.

6 more elements

A system for detecting features of a character stored within a binary memory, said binary memory comprising a plurality of cells with at least some cells containing binary information representing said character, said character having at least one boundary, the one boundary comprising an outer boundary, any other boundaries being inner loops, comprising:

(a) scanning means connected to said memory for locating at least one cell in said memory containing information on each boundary;
(b) vector means connected to produce vectors representative of the slope between adjacent cells containing information concerning said character located on each boundary by tracing each boundary in response to said scanning means locating said one cell thereof;
(c) averaging means connected to average a certain number of consecutive vectors received from said vector means to form a set of vector averages for each boundary and;
(d) inflection means connected to compare consecutive vector averages of said set for locating inflection points from differences between consecutive vector averages of said set by signaling when a difference of one sign is followed by a difference of opposite sign.

1 more element

Dependencies are inferred from the claim text.

Slope feature detection system

Life of this patent

Slope feature detection system

Life of this patent

Abstract

Claims14