EP0689153A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Character recognition
Assignee: Texas Instruments Italia SRL
•
Filing: 1995-05-26
•
Granted: —
•
Expires: 2015-05-26
Abstract
Abstract This invention relates to a method and apparatus for recognizing a script written character. The character is entered using character entering means and digitised by appropriate means. The digitised character is then stored in, for example, a memory. Codes representing topological, vector dimension features and the microfeatures of the character are extracted from the character, then the features of the character are compared with a set of reference features corresponding thereto stored in a memory. Each of the set of reference characters corresponding with a specific script written character. A logic process is then performed to determine which of the set of reference features most closely corresponds to the topological features of the digitized character thereby identifying the script written character. The relative weighting of the feature can be varied for different types of script or confusing characters to enable still more accurate recognition.
Inventors
- 👤 Girolamo Texas Instruments Italia Spa Gallo
- 👤 Cristina Texas Instruments Italia Spa Lattaro
- 👤 Giuseppe Texas Instruments Italia Spa Savarese
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-09 11:33
CPC Classifications
G; G06; G06V; G06V30/00; G06V30/10; G06V30/18; G06V30/1801; G06V30/18019
Claims Analysis
Total Claims: 21
Claim 1
Independent Claim
118 words
A method for recognising a script written character comprising the steps of entering the character using character entering means; digitizing the character; storing the digitized character; extracting topological features and vector features of said character; extracting microfeatures of the character; determining dimensions of the character; comparing the topological and vector features, microfeatures and dimensions of the character with a set of reference features equivalent thereto and stored in a memory, each of the set corresponding to a specific character; and performing a logic process to determine which of the set of reference features most closely corresponds to the topological features, the vector features, the microfeatures and the dimensions of the digitized character, thereby recognising the script written character.
Claim 2
Dependent Claim
41 words
The method of claim 1, wherein extracting the features comprises weighting the features of one of the topology, the vector, the microfeature and the dimensions more than the others in a predetermined manner in dependance on the nature of the character.
Claim 3
Dependent Claim
30 words
The method of claim 1 or claim 2, wherein extracting microfeatures of the character comprising obtaining optical information relating to the nature of the character to produce of microfeature code.
Claim 4
Independent Claim
29 words
The method of any preceding claim, wherein digitizing the character comprises constructing a matrix in ones and zeros and an array of points which are representative of the character.
Claim 5
Dependent Claim
47 words
The method of claim 4, further comprising scanning the matrix representing the character to determine a microfeature code when a black pixel is found in a row of the matrix memorising the neighbouring pixels in the row below and assigning a code indicative of the neighbouring pixels.
Claim 6
Dependent Claim
70 words
The method of claim 3 and any of claims 4 or 5 when dependant on claim 3, wherein the microfeature code is given by: where: Rik (k=1,....5, i=1.....15) is the microfeature code of the input character, Rjik (k=1,....5, i=1.....15) is the microfeature code of the j-th reference character, P is the number of non-zero Rik elements, Pj is the number of non-zero Rjik elements, and M is the following matrix:
Claim 7
Dependent Claim
25 words
The method of claim 5 or 6, wherein extracting the vector code comprises extracting a code indicative of the number of points of said character.
Claim 8
Independent Claim
34 words
The method of any of claims 4 to 7, wherein extracting the vector code further comprises extracting a code which describes the position of each point of the array relative to the previous point.
Claim 9
Independent Claim
53 words
The method of any preceding claim, wherein the step of determining dimensions of the character comprises determining X and Y dimensions of the character and determining a dynamic score and performing a weighting operation on each of the X and Y dimensions to enable differentiation of characters have similar shapes, but different dimensions.
Claim 10
Independent Claim
35 words
The method of any preceding claim, wherein the entering step comprises inputting the character into a frame on the frame and entering means assigning codes to the character dependant on its position in that frame.
Claim 11
Independent Claim
160 words
The method of any preceding claim, wherein performing the logic process comprises evaluating a score Sj for each of the reference characters (Ncar) and identifying the reference character with the lowest score Sj as the character that was entered as the script written character, wherein Sj is given by: Si=Wf*Sfj-Wm*Smj-Wd*Sdj+Ws*|St-Stj|+Wx*|X-Xj|+Wy*|Y-Yj| where: Sfj is the feature extraction score Smj is the microfeature detection score Sdj is the dynamic score Wf is a weighting factor for Sfj Wm is a weighting factor for Smj Wd is a weighting factor for Sdj St is the stroke number of the input character Stj is the stroke number of the j-th reference character Ws is the stroke weighting factor X is the width of the input character Xj is the width of the j-th reference character Wx is the width weighting factor Y is the height of the input character Yj is the height of the j-th reference character Wy is a height weighting factor
Claim 12
Dependent Claim
104 words
Apparatus for recognising a script written character comprising: character enter means for entering and receiving the character; digitizing means for digitizing the character; storage means for storing the digitized character; feature extracting means for extracting topological features; vector features microfeatures and dimension features representative of said character; comparing means for comparing the extracted features of the character with a set of reference features stored in a memory, each of the set corresponding to a specific character; and logic means for identifying which of the set of reference features most closely corresponds to the features of the digitized character, thereby recognising the script written character.
Claim 13
Dependent Claim
41 words
Apparatus according to claim 12, further including weighting means associated with the feature extracting means capable of giving a different weighting factor to each of said features in a predetermined manner which is in dependance of the nature of the character.
Claim 14
Dependent Claim
22 words
Apparatus according to claim 13, wherein the vector features weighting factor is higher than the topological features weighting factor for Japanese script.
Claim 15
Dependent Claim
22 words
Apparatus according to claim 13, wherein the vector features weighting factor is lower than the topological features weighting factor for Roman script.
Claim 16
Dependent Claim
24 words
Apparatus according to claim 13, wherein the microfeatures and the dimension features are used to resolve the identity of characters which have similar characteristics.
Claim 17
Dependent Claim
28 words
Apparatus according to any of claims 12 to 16, wherein the feature extracting means comprise code generating means for generating a plurality of codes indicative of the features.
Claim 18
Dependent Claim
25 words
Apparatus according to any of claims 12 to 17, wherein the comparing means compares codes of the reference set and code for the input character.
Claim 19
Dependent Claim
120 words
Apparatus according to any of claims 12 to 18, wherein the logic means evaluated a score Sj for each reference character (Nd) and identifies the reference character with the lowest score as that corresponding to the script written characters, wherein Sj is given by: where: S = stroke number of the input character Stj = stroke number of the j-th reference character fi = response of the i-th feature to the input character fij = response of the i-th feature to the j-th reference character Ni = number of intersections of the i-th line with the input character Nij = number of intersections of the i-th line with the j-th reference character SW = stroke weight IW = intersection weight
Claim 20
Dependent Claim
40 words
Apparatus according to any of claims 12 to 19, wherein an interface is provided between the apparatus and a computer such that the interface passes the identified character to the computer thereby forming part of the instructions for the computer.
Claim 21
Dependent Claim
32 words
A computer based product including apparatus according to any of claims 12 to 20, and an interface between the apparatus and the computer for passing instruction to the computer from the apparatus.
References Cited (Backward Citations)
US4972496A
Cited
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.