Systems and methods for capturing handwritten information using handwriting analysis

• US 7,272,259 B2
• Filed: 12/12/2003
• Issued: 09/18/2007
• Est. Priority Date: 03/15/2002
• Status: Active Grant

• Current Assignee: Computer Sciences Corporation (DXC Technology Company)
• Original Assignee: Computer Sciences Corporation (DXC Technology Company)
• Inventors: Mason, James G., Blinova, Katerina, Berkhuysen, Johan Willem Piere, Shridhar, Malayappan, Nugmanov, Babur, Bakker, Ronny, Houle, Gilles
• Primary Examiner(s): CHAWAN, SHEELA C
• Application Number:
  US10/734,738
• Publication Number:
  US 20040228514A1
• Time in Patent Office: 1,376 Days
• Field of Search: 382/187, 382/190, 382/168, 382/176, 382/177, 382/173, 382/181, 382/179, 382/188, 382/206, 382/232, 382/254, 382/278
• US Class Current: 382/187
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/133 Evaluation of quality of the acquired characters
• CPC Class Codes: G06V 30/274 Syntactic or semantic context, e.g. balancing
• CPC Class Codes: G06V 40/30 Writer recognition; Reading and verifying signatures

Abstract

Methods and systems are provided for analyzing and assessing documents using a writing profile for documents, such as a payment instrument. A method may include providing a document to a computer system. In an embodiment, the method may further include assessing whether writing in an information field approximately matches a writing profile representation from at least one information field from at least one other document. In certain embodiments, the matching writing profile representation may be associated with a corresponding text representation in a computer processable format in memory on the computer system. In some embodiments, the information field may be associated with the text representation corresponding to the matching writing profile representation.

Claims

1. A method of assessing information in a document using a computer system, comprising:

providing the document to the computer system, wherein the document comprises at least two information fields;

obtaining information on writing in an information field of the document; and

comparing the obtained written information in the information field and written information in at least one other information field to at least one writing profile representation from at least one other document,

wherein at least one of the writing profile representations comprises written information from the information field and written information from at least one of the other information fields from at least one of the other documents, wherein at least one writing profile representation comprises at least one cross-field correlation between at least two information fields.

2. The method of claim 1, further comprising using at least one of the comparisons of written information to verify the obtained written information.

3. The method of claim 1, further comprising assessing whether the obtained written information in the information field and the written information in at least one other information field approximately matches at least one writing profile representation from at least one other document.

4. The method of claim 1, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least the one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field.

5. The method of claim 1, wherein obtaining information on writing comprises recognizing written information.

6. The method of claim 1, further comprising assessing a frequency associated with the written information in the information field if the written information in the information field and written information in at least one other information field approximately matches at least one writing profile representation.

7. The method of claim 1, further comprising assessing whether the written information in the information field comprises written information associated with a particular frequency.

8. The method of claim 1, wherein providing the document to the computer system comprises providing images of the document to the computer system.

9. The method of claim 1, wherein comparing written information comprises comparing at least one characteristic of the writing.

10. The method of claim 1, wherein at least handwriting profile representation is obtained from a valid document.

11. The method of claim 1, wherein at least one of the documents comprises a payment instrument, and wherein the writing is writing from at least one account owner of an account of the payment instrument.

12. The method of claim 1, wherein providing the document to the computer system comprises obtaining images of writing of at least one information field.

13. The method of claim 1, wherein the written information in the information field comprises an image.

14. The method of claim 1, wherein the written information comprises at least one mathematical representation.

15. The method of claim 1, wherein the written information in the information field comprises at least one image.

16. The method of claim 1, wherein the written information in the information field comprises at least one type of written information.

17. The method of claim 1, wherein the written information in the information field comprises at least one type of written information, and wherein at least one type of written information comprises a word type.

18. The method of claim 1, wherein the written information in the information field comprises at least one type of written information, and wherein at least one type of written information comprises a character type.

19. The method of claim 1, wherein the written information in the information field comprises at least one global feature of the writing.

20. The method of claim 1, wherein the written information in the information field comprises at least one local feature of the writing.

21. The method of claim 1, wherein the written information in the information field comprises at least one syntax pattern.

22. The method of claim 1, wherein the written information in the information field comprises at least one lexicon name for at least one information field.

23. The method of claim 1, wherein at least one writing profile representation comprises at least one mathematical representation.

24. The method of claim 1, wherein at least one writing profile representation comprises at least one image.

25. The method of claim 1, wherein at least one writing profile representation comprises at least one type of written information.

26. The method of claim 1, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information.

27. The method of claim 1, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information, and wherein at least one type of written information comprises a word type.

28. The method of claim 1, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information, and wherein at least one type of written information comprises a character type.

29. The method of claim 1, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information, and further comprising determining at least one of the variants with a cluster algorithm.

30. The method of claim 1, wherein at least one writing profile representation comprises at least one global characteristic of the writing.

31. The method of claim 1, wherein at least one writing profile representation comprises at least one local characteristic of the writing.

32. The method of claim 1, wherein at least one writing profile representation comprises at least one variant of a syntax pattern.

33. The method of claim 1, wherein at least one writing profile representation comprises at least one lexicon name for at least one information field.

34. The method of claim 1, wherein the document comprises a payment instrument, wherein a the information field comprises a signature field of the payment instrument, wherein at least one of the other information fields is a non-signature field of the payment instrument.

35. The method of claim 34, wherein the at least one other information field comprises an amount field of the payment instrument.

36. The method of claim 34, wherein the at least one other information field comprises a payee field of the payment instrument.

37. The method of claim 34, wherein the at least one other information field comprises a date field of the payment instrument.

38. The method of claim 1, wherein the document comprises a payment instrument, wherein the written information being compared comprises written information from at least two non-signature fields.

39. The method of claim 1, wherein the at least one other information field comprises pre-printed information in the document.

40. The method of claim 1, wherein the document comprises a payment instrument, wherein comparing the obtained written information in the information field and written information in at least one other information field comprises comparing the pre-printed information against a stock profile representation associated with at least one other document and assessing potential fraud based on the comparison of the pre-printed information against the stock profile representation.

41. The method of claim 40, wherein the document comprises a payment instrument, wherein comparing pre-printed information in the document against a stock profile representation comprises comparing pre-printed information against layout characteristics of the stock profile representation.

42. The method of claim 40, wherein the document comprises a payment instrument, wherein comparing pre-printed information in the document against a stock profile representation comprises comparing pre-printed information against characteristics of graphic elements of the stock profile representation.

43. The method of claim 1, further comprising performing one or more fraud tests and assigning a fraud weight to at least one of the one or more of the fraud tests, wherein the fraud weight corresponds to the strength of the indication of fraud in the payment instrument.

44. The method of claim 1, wherein comparing the obtained written information in the information field and written information in at least one other information field to at least one writing profile representation from at least one other document comprises:

performing at least two fraud tests on at least one of the information fields of the document.

45. The method of claim 44, wherein one of the fraud tests is weighted differently than at least one of the other fraud tests.

46. The method of claim 44, wherein a fraud test for one of the information fields is weighted higher than a fraud test for another of the information fields.

47. The method of claim 44, wherein a fraud test for a frequently present feature of a payment instrument are weighted higher than a fraud test involving an infrequently present feature of the payment instrument.

48. The method of claim 44, wherein a fraud test for at least one an amount field of a payment instrument is weighted higher than a fraud test for at least one other information field of the payment instrument.

49. The method of claim 1, wherein comparing the obtained written information in the information field and written information in at least one other information field to at least one writing profile representation from at least one other document comprises:

assessing whether written information in a first information field of the document approximately matches written information from a corresponding information field the at least one other document; and

assessing whether written information in a second information field of the document approximately matches written information from a corresponding information field in the at least one other document.

50. The method of claim 49, wherein comparing the obtained written information in the information field and written information in at least one other information field to at least one writing profile representation from at least one other document comprises comparing characteristics or patterns of at least one of the information fields in the document with characteristics or patterns of a corresponding information field in the at least one other document.

51. The method of claim 49, wherein comparing the obtained written information in the information field and written information in at least one other information field to at least one writing profile representation from at least one other document comprises comparing data content of at least one of the information fields in the document with data content of a corresponding information field in the at least one other document.

52. The method of claim 1, further comprising analyzing a correlation of information between two or more of the information fields.

53. The method of claim 1, wherein at least one writing profile representation comprises at least one lexicon name for at least one information field.

54. The method of claim 1, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least the one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field.

55. The method of claim 1, wherein at least one writing profile representation comprises at least one cross-field correlation between a payee name field and at least one other field.

56. The method of claim 1, further comprising analyzing a correlation of information between two or more non-signature information fields.

57. The method of claim 56, wherein one of the non-signature information fields is a payee name field.

58. The method of claim 56, wherein one of the non-signature information fields is an amount field.

59. The method of claim 1, the method further comprising analyzing a correlation of information between two or more of the information fields, wherein the analysis of a first information field of the document is dependent on what information is in a second information field of the document.

60. The method of claim 1, further comprising predicting a likely entry type in one information field of the document based on an entry in another of the information fields of the document.

61. The method of claim 1, further comprising assessing whether written information in a first information field of the document was written by the same person as written information in at least one other information field of the document.

62. The method of claim 61, wherein assessing whether written information in the first information field of the document is written by the same person as at least one other information field of the document comprises determining if there is an approximate match of at least one global characteristic in the first information field of the document and at least one of the other information fields of the document.

63. The method of claim 62, wherein the first information field and at least one of the other information fields are non-signature fields.

64. A system, comprising:

a CPU;

a data memory coupled to the CPU; and

a system memory coupled to the CPU, wherein the system memory is configured to store one or more computer programs executable by the CPU, and wherein the computer programs are executable to implement a method of assessing information in a document, the method comprising;

providing the document to the computer system, wherein the document comprises at least two information fields;

obtaining information on writing in an information field of the document; and

comparing the obtained written information in the information field and written information in at least one other information field to at least one writing profile representation from at least one other document,

wherein at least one of the writing profile representations comprises written information from the information field and written information from at least one of the other information fields from at least one of the other documents, wherein at least one writing profile representation comprises at least one cross-field correlation between at least two information fields.

65. The system of claim 64, further comprising using at least one of the comparisons of written information to verify the obtained written information.

66. The system of claim 64, further comprising assessing whether the obtained written information in the information field and the written information in at least one other information field approximately matches at least one writing profile representation from at least one other document.

67. The system of claim 64, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least the one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field.

68. The system of claim 64, wherein obtaining information on writing comprises recognizing written information.

69. A computer readable medium comprising program instructions, wherein the program instructions are computer-executable to implement a method of assessing information in a document, the method comprising:

providing the document to the computer system, wherein the document comprises at least two information fields;

obtaining information on writing in an information field of the document; and

comparing the obtained written information in the information field and written information in at least one other information field to at least one writing profile representation from at least one other document,

wherein at least one of the writing profile representations comprise written information from the information field and written information from at least one of the other information fields from at least one of the other documents, wherein at least one writing profile representation comprises at least one cross-field correlation between at least two information fields.

70. The computer readable medium of claim 69, further comprising using at least one of the comparisons of written information to verify the obtained written information.

71. The computer readable medium of claim 69, further comprising assessing whether the obtained written information in the information field and the written information in at least one other information field approximately matches at least one writing profile representation from at least one other document.

72. The computer readable medium of claim 69, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least the one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field.

73. The computer readable medium of claim 69, wherein obtaining information on writing comprises recognizing written information.

74. A method of assessing information in a document using a computer system, comprising:

providing the document to the computer system, wherein the document comprises at least two information fields;

obtaining information on writing in an information field of the document;

comparing the obtained written information in the information field and written information in at least two other information fields to at least one writing profile representation from at least one other document,

wherein at least one of the writing profile representations comprise written information from the information field and written information from at least two of the other information fields from at least the one of the other documents, wherein at least one writing profile representation comprises at least one cross-field correlation between at least two information fields.

75. The method of claim 74, further comprising using at least one of the comparisons of written information to verify the obtained written information.

76. The method of claim 74, further comprising assessing whether the obtained written information in the information field and the written information in at least two other information fields approximately matches at least one writing profile representation from at least one other document.

77. The method of claim 74, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least the one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field.

78. The method of claim 74, wherein obtaining information on writing comprises recognizing written information.

79. The method of claim 74, further comprising assessing a frequency associated with the written information in the information field if the written information in the information field and written information in at least one other information field approximately matches at least one writing profile representation.

80. The method of claim 74, further comprising assessing whether the written information in the information field comprises written information associated with a particular frequency.

81. The method of claim 74, wherein providing the document to the computer system comprises providing images of the document to the computer system.

82. The method of claim 74, wherein comparing written information comprises comparing at least one characteristic of the writing.

83. The method of claim 74, wherein at least handwriting profile representation is obtained from a valid document.

84. The method of claim 74, wherein at least one of the documents comprises a payment instrument, and wherein the writing is writing from at least one account owner of an account of the payment instrument.

85. The method of claim 74, wherein providing the document to the computer system comprises obtaining images of writing of at least one information field.

86. The method of claim 74, wherein the written information in the information field comprises an image.

87. The method of claim 74, wherein the written information comprises at least one mathematical representation.

88. The method of claim 74, wherein the written information in the information field comprises at least one image.

89. The method of claim 74, wherein the written information in the information field comprises at least one type of written information.

90. The method of claim 74, wherein the written information in the information field comprises at least one type of written information, and wherein at least one type of written information comprises a word type.

91. The method of claim 74, wherein the written information in the information field comprises at least one type of written information, and wherein at least one type of written information comprises a character type.

92. The method of claim 74, wherein the written information in the information field comprises at least one global feature of the writing.

93. The method of claim 74, wherein the written information in the information field comprises at least one local feature of the writing.

94. The method of claim 74, wherein the written information in the information field comprises at least one syntax pattern.

95. The method of claim 74, wherein the written information in the information field comprises at least one lexicon name for at least one information field.

96. The method of claim 74, wherein at least one writing profile representation comprises at least one mathematical representation.

97. The method of claim 74, wherein at least one writing profile representation comprises at least one image.

98. The method of claim 74, wherein at least one writing profile representation comprises at least one type of written information.

99. The method of claim 74, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information.

100. The method of claim 74, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information, and wherein at least one type of written information comprises a word type.

101. The method of claim 74, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information, and wherein at least one type of written information comprises a character type.

102. The method of claim 74, wherein at least one writing profile representation comprises at least one writing variant of an example of at least one type of written information, and further comprising determining at least one of the variants with a cluster algorithm.

103. The method of claim 74, wherein at least one writing profile representation comprises at least one global characteristic of the writing.

104. The method of claim 74, wherein at least one writing profile representation comprises at least one local characteristic of the writing.

105. The method of claim 74, wherein at least one writing profile representation comprises at least one variant of a syntax pattern.

106. The method of claim 74, wherein at least one writing profile representation comprises at least one lexicon name for at least one information field.

107. The method of claim 74, wherein the document comprises a payment instrument, wherein a the information field comprises a signature field of the payment instrument, wherein at least one of the other information fields is a non-signature field of the payment instrument.

108. The method of claim 107, wherein the at least two other information fields comprise an amount field of the payment instrument.

109. The method of claim 107, wherein the at least two other information fields comprise a payee field of the payment instrument.

110. The method of claim 107, wherein the at least two other information fields comprise a date field of the payment instrument.

111. The method of claim 74, wherein the document comprises a payment instrument, wherein the written information being compared comprises written information from at least two non-signature fields.

112. The method of claim 74, wherein the at least one other information field comprises pre-printed information in the document.

113. The method of claim 74, wherein the document comprises a payment instrument, wherein comparing the obtained written information in the information field and written information in at least two other information fields comprises comparing the pre-printed information against a stock profile representation associated with at least one other document and assessing potential fraud based on the comparison of the pre-printed information against the stock profile representation.

114. The method of claim 113, wherein the document comprises a payment instrument, wherein comparing pre-printed information in the document against a stock profile representation comprises comparing pre-printed information against layout characteristics of the stock profile representation.

115. The method of claim 113, wherein the document comprises a payment instrument, wherein comparing pre-printed information in the document against a stock profile representation comprises comparing pre-printed information against characteristics of graphic elements of the stock profile representation.

116. The method of claim 74, further comprising performing one or more fraud tests and assigning a fraud weight to at least one of the one or more of the fraud tests, wherein the fraud weight corresponds to the strength of the indication of fraud in the payment instrument.

117. The method of claim 74, wherein comparing the obtained written information in the information field and written information in at least two other information fields to at least one writing profile representation from at least one other document comprises:

performing at least two fraud tests on at least one of the information fields of the document.

118. The method of claim 117, wherein one of the fraud tests is weighted differently than at least one of the other fraud tests.

119. The method of claim 117, wherein a fraud test for one of the information fields is weighted higher than a fraud test for another of the information fields.

120. The method of claim 117, wherein a fraud test for a frequently present feature of a payment instrument are weighted higher than a fraud test involving an infrequently present feature of the payment instrument.

121. The method of claim 117, wherein a fraud test for at least one an amount field of a payment instrument is weighted higher than a fraud test for at least one other information field of the payment instrument.

122. The method of claim 74, wherein comparing the obtained written information in the information field and written information in at least two other information fields to at least one writing profile representation from at least one other document comprises:

assessing whether written information in a first information field of the document approximately matches written information from a corresponding information field the at least one other document;

assessing whether written information in a second information field of the document approximately matches written information from a corresponding information field the at least one other document; and

assessing whether written information in a third information field of the document approximately matches written information from a corresponding information field in the at least one other document.

123. The method of claim 122, wherein comparing the obtained written information in the information field and written information in at least two other information fields to at least one writing profile representation from at least one other document comprises comparing characteristics or patterns of at least one of the information fields in the document with characteristics or patterns a corresponding information field in the at least one other document.

124. The method of claim 122, wherein comparing the obtained written information in the information field and written information in at least two other information fields to at least one writing profile representation from at least one other document comprises comparing data content of at least one of the information fields in the document with data content of a corresponding information field in the at least one other document.

125. The method of claim 74, further comprising analyzing a correlation of information between two or more of the information fields.

126. The method of claim 74, wherein at least one writing profile representation comprises at least one lexicon name for at least one information field.

127. The method of claim 74, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least the one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field.

128. The method of claim 74, wherein at least one writing profile representation comprises at least one cross-field correlation between a payee name field and at least one other field.

129. The method of claim 74, further comprising analyzing a correlation of information between two or more non-signature information fields.

130. The method of claim 129, wherein one of the non-signature information fields is a payee name field.

131. The method of claim 130, wherein one of the non-signature information fields is an amount field.

132. The method of claim 74, the method further comprising analyzing a correlation of information between two or more of the information fields, wherein the analysis of a first information field of the document is dependent on what information is in a second information field of the document.

133. The method of claim 74, further comprising predicting a likely entry type in one information field of the document based on an entry in another of the information fields of the document.

134. The method of claim 74, further comprising assessing whether written information in a first information field of the document was written by the same person as written information in at least one other information field of the document.

135. The method of claim 134, wherein assessing whether written information in the first information field of the document is written by the same person as at least one other information field of the document comprises determining if there is an approximate match of at least one global characteristic in the first information field of the document and at least one of the other information fields of the document.

136. The method of claim 135, wherein the first information field and at least one of the other information fields are non-signature fields.

137. A system, comprising:

a CPU;

a data memory coupled to the CPU; and

a system memory coupled to the CPU, wherein the system memory is configured to store one or more computer programs executable by the CPU, and wherein the computer programs are executable to implement a method of assessing information in a document, the method comprising;

providing the document to the computer system, wherein the document comprises at least two information fields;

obtaining information on writing in an information field of the document;

comparing the obtained written information in the information field and written information in at least two other information fields to at least one writing profile representation from at least one other document,

wherein at least one of the writing profile representations comprise written information from the information field and written information from at least two of the other information fields from at least the one of the other documents, wherein at least one writing profile representation comprises at least one cross-field correlation between at least two information fields.

138. The system of claim 137, further comprising using at least one of the comparisons of written information to verify the obtained written information.

139. The system of claim 137, further comprising assessing whether the obtained written information in the information field and the written information in at least one other information field approximately matches at least one writing profile representation from at least one other document.

140. The system of claim 137, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field.

141. The system of claim 137, wherein obtaining information on writing comprises recognizing written information.

142. A computer readable medium comprising program instructions, wherein the program instructions are computer-executable to implement a method of assessing information in a document, the method comprising:

providing the document to the computer system, wherein the document comprises at least two information fields;

obtaining information on writing in an information field of the document;

comparing the obtained written information in the information field and written information in at least two other information fields to at least one writing profile representation from at least one other document,

wherein at least one of the writing profile representations comprise written information from the information field and written information from at least two of the other information fields from at least the one of the other documents, wherein at least one writing profile representation comprises at least one cross-field correlation between at least two information fields.

143. The computer readable medium of claim 142, further comprising using at least one of the comparisons of written information to verify the obtained written information.

144. The computer readable medium of claim 142, further comprising assessing whether the obtained written information in the information field and the written information in at least one other information field approximately matches at least one writing profile representation from at least one other document.

145. The computer readable medium of claim 142, wherein at least one writing profile representation comprises a cross-field correlation between the information field and at least the one other information field, wherein the cross-field correlation comprises a lexicon name for the information field and at least one lexicon name for at least the one other information field. writing comprises recognizing written information.

146. The computer readable medium of claim 142, wherein obtaining information on writting comprises recognizing written information.