DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Typographic Conventions
Throughout this office action, shorthand notation for referencing locations of elements in documents are utilized. The following is a brief summary of the shorthand utilized:
Sec. – is used to denote an associated section with a header in non-patent literature
¶ – is used to denote the number and location of a paragraph
col. – is used to denote a column number
ln. – is used to denote a line; if a line number is not demarcated in a document, the line number will be assumed to start at 1 for each paragraph.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 04/11/2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Objections
Claims 7 & 14 are objected to because of the following informalities:
Claim 7 recites “target document in an array in same arrangement”. The examiner believes this was intended to read as “target document in an array in the same arrangement”
Claim 14 recites “the frequently occurring word string of each of the predetermined document types being not the overlapping frequently occurring word string”. While the examiner notes that phrase is intended to restrict the selection of a document type to not factor in the spatial relationships of frequently occurring word strings appearing across multiple predetermined types, the syntax this phrase is difficult to interpret.
Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-15, & 19 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 1, 13 & 19 recite the limitation “information indicating appropriateness of the document being a document of the predetermined document type is output”. The specification further details this appropriateness (expressed as reliability or likelihood) of the document being a probability of the document being the predetermined document type. The specification, however, fails to disclose the bounds of what would be considered “appropriate” with respect to the claim language. The term “appropriateness” is indefinite for lacking a proper definition or threshold.
According to MPEP § 2173.05(b), a claim is held indefinite when “the specification lacked some standard for measuring the degrees intended”. It is apparent that the claim limitation for “appropriateness” within the context of the claim and specification unclear to one of ordinary skill in the art.
Therefore, claims 1, 13, & 19 are rejected as indefinite for the claim limitation “appropriateness”.
Claims 2-15 are rejected as being dependent on claim 1.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-8, 13, 15-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Sampson et al (US 8724907 B1).
Regarding claim 1, Sampson et al disclose a system and method for classifying documents using optical character recognition (OCR) data. More specifically, Sampson et al teach An information processing system (system 405 for grouping and classifying documents [col. 4 | ln. 50-53; Fig. 4]) comprising:
circuitry (central processor 302 [col. 3 | ln. 53-55; Fig. 3]) configured to:
acquire a character recognition result of an identification target image that is an image of an identification target document (receive OCR data of a document image [col. 7 | ln. 8-13]);
store a frequently occurring word string of a predetermined document type (in step 1710, the system creates and stores templates corresponding to each document class, with each template including a set of keywords [col. 20 | ln. 37-43; Fig. 17] – the examiner notes that the “frequently occurring word string” is being interpreted as a word that is frequently occurring across different documents of the same type, in this case, a keyword or set of keywords.);
detect the frequently occurring word string from the character recognition result of the identification target image to acquire information on a position of the frequently occurring word string in the identification target document (in steps 1715 & 1720, the system receives an input document for classification and compares the spatial relationship of keywords to other words in the document based on their location in the document [col. 20 | ln. 44-56; Fig. 17]);
generate a feature quantity of the identification target document using the information on the position, the feature quantity including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string and another word string in the identification target document (the system obtains information regarding the location of both a keyword and a different word in the document to be compared, and compares it to their relationships in the template document [col. 20 | ln. 49-54]; this is further exemplified in Fig. 14, illustrating the spatial relationship of words across two documents [col .14 | ln. 1-9]);
store a trained model that identifies the predetermined document type (system 405 comprises a training module 410, a classification module 415, and a word location comparison engine 420, which is used to group and classify documents [col. 4 | ln. 50-54; Fig. 4]), the trained model being generated through machine learning (the system 405 includes a training module 415 that receives a set of training documents 425 to output a set of document classes 430 and associated document templates 435, which can be generated via automatic training [col. 4 | ln. 52-67 & col. 5 | ln. 1 – 15; Fig. 4], and that during the automatic training step, a distance function is used to determine whether a document is “close” to a template [col. 5 | ln. 62 – 65], which is further outlined in Fig. 9, with step 920 outlining a clustering algorithm to compare and group similar documents [col. 8 | ln. 1-12] – the examiner notes that that “automatic training” of the training model is interpreted as machine learning.) such that, in response to input of a feature quantity of a document including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string and another word string in the document (a comparison function takes two images and finds a set of common words in approximately the same relative position in each document [col. 8 | ln. 23-30], with step 1720 comparing the position of keywords with other words in the document [col. 20 | ln. 49-54]), information indicating appropriateness of the document being a document of the predetermined document type is output (a scoring function provides a score proportional to a number and size of the common word across each document [col. 8 | ln. 29-33]); and
input the feature quantity of the identification target document to the trained model to identify whether the identification target document is a document of the predetermined document type (in step 1725, the spatial relationship between a keyword and another word is provided to the training model to compare to a template and classify the document according to that template [col. 20 | ln. 49-54; Fig. 17]).
Regarding claim 2, Sampson et al teach The information processing system of claim 1 (as described above), wherein the trained model is generated through machine learning using training data (the system 405 includes a training module 415 that receives a set of training documents 425 to output a set of document classes 430 and associated document templates 435, which can be generated via automatic training [col. 4 | ln. 52-67 & col. 5 | ln. 1 – 15; Fig. 4], and that during the automatic training step, a distance function is used to determine whether a document is “close” to a template [col. 5 | ln. 62 – 65], which is further outlined in Fig. 9, with step 920 outlining a clustering algorithm to compare and group similar documents [col. 8 | ln. 1-12]),
the training data associating, for each of a plurality of training images including a plurality of predetermined document type images that are images of documents of the predetermined document type having layouts different from one another (a class of documents is derived from a set of training documents of a known predetermined type in step 905 [col. 6 | ln. 49-51; Fig. 9], with Figs. 6 & 7 illustrating two documents of the same type (in this case, an invoice, with different layouts, but the same keywords [col. 6 | ln. 21-48]), a feature quantity of a document depicted in the training image with information indicating whether the document depicted in the training image is a document of the predetermined document type (during training, a location comparison engine 420 compares documents based on the position of keyword and its spatial relation relative to other words (e.g., a feature quantity) to determine whether documents are of the same class [col. 4 | ln. 65-67 & col. 5 | ln. 1-43]),
the feature quantity of a document depicted in the training image including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string and another word string in the document depicted in the training image (during training, a location comparison engine 420 compares documents based on the position of keyword and its spatial relation relative to other words (e.g., a feature quantity) to determine whether documents are of the same class [col. 4 | ln. 65-67 & col. 5 | ln. 1-43]).
Regarding claim 3, Sampson et al teach The information processing system of claim 1 (as previously described), wherein
the frequently occurring word string is one of a plurality of frequently occurring word strings (the system has a list of keywords which are considered significant to that reference type [col. 16 | ln. 10 – 23]), and
the positional relationship feature quantity includes a feature quantity indicating a distance between the frequently occurring word string and another frequently occurring word string in the identification target document (Fig. 14 illustrates the positional relationship of commonly occurring words relative to each other across two documents, with exemplary vectors (like 1416a-c) representing textual distances from pivot word1 (1415) [col 14 | ln. 1-62]).
Regarding claim 4, Sampson et al teach The information processing system of claim 1 (as previously described), wherein the positional relationship feature quantity includes a feature quantity indicating a size of a row including the frequently occurring word string (the document can be partitioned into a grid based on a predefined threshold, resulting in rows directly proportional to the grid size, which indicate a size of a row as well as a position (which row) a word appears in [col. 12 | ln. 1-52; Fig. 12 & 13]).
Regarding claim 5, Sampson et al teach The information processing system of claim 1 (as previously described), wherein the feature quantity of the identification target document includes the positional relationship feature quantity and a feature quantity indicating an attribute of the frequently occurring word string (the examiner notes that the “attribute” is being interpreted as a position or size of the word in a document, in accordance with the specification of the present application [pg. 14 | ln. 32-34] – the comparison function can find a set of common words, taking into account both their size or positions in the document [col. 8 | ln. 22-35]).
Regarding claim 6, --Sampson et al teach The information processing system of claim 5 (as previously described), wherein the feature quantity indicating the attribute of the frequently occurring word string includes at least one of a feature quantity indicating a position of the frequently occurring word string or (given the use of the disjunctive “or” only one of either listed claim limitation is needed for mapping) a feature quantity indicating a size of the frequently occurring word string (the comparison function can find a set of common words, taking into account both their size or positions in the document [col. 8 | ln. 22-35]).
Regarding claim 7, Sampson et al teach The information processing system of claim 2 (as previously described), wherein the circuitry (central processor 302 [col. 3 | ln. 53-55; Fig. 3]) is configured to:
store the trained model generated using training data (system 405 comprises a training module 410, stored in the system memory 304 [col. 3 | ln. 45-55 & col. 4 | ln. 50-54; Figs. 3 & 4]),
the training data associating a feature array with information indicating whether the document depicted in each of the plurality of training images is a document of the predetermined document type (using the training documents 425, the training module 410 outputs a set of document classes 430 and associated document templates 435 for a given document class [col. 4 | ln. 52-64]), the feature array having the feature quantity of the document depicted in each of the plurality of training images been aggregated in an array form (the document templates 435 of a known class contain keywords and their spatial relationships to other words in the document [ col. 5 | ln. 28-43]);
form the feature quantity of the identification target document in an array in same arrangement order as the feature array (in step 1720, the spatial relationship of keywords relative to other words in a document is compared to the spatial relationship of keywords relative to other words in a training document [col. 20 | ln. 37-51; Fig. 17]); and
input, to the trained model, the feature quantity of the identification target document formed in the array, to identify whether the identification target document is a document of the predetermined document type (in step 1725, the document is subsequently classified based on the comparison [col. 20 | ln. 51-56; Fig. 17]).
Regarding claim 8, Sampson et al teach The information processing system of claim 1 (as previously described), wherein
the predetermined document type is one of a plurality of predetermined document types (a set of document classes 430 and associated document template 435 is output by training module 425 [col. 4 | ln. 50-54; Fig. 4]), and
the circuitry (central processor 302 [col. 3 | ln. 53-55; Fig. 3]) is configured to:
store, for each of the plurality of predetermined document types, a trained model that identifies the predetermined document type (a training module 425 is stored in system 405 to inform a word location comparison engine 420 to compare documents to be classified 440 by the classification module 415 [col. 4 | ln. 50-54; Fig. 4]);
for each of the plurality of predetermined document types, identify whether the identification target image corresponds to the predetermined document type using the trained model that identifies the predetermined document type (in step 1720, the spatial relationship of keywords relative to other words in a document is compared to the spatial relationship of keywords relative to other words in a training document [col. 20 | ln. 37-51; Fig. 17]); and
identify, based on a result of the identification for each of the plurality of predetermined document types, which document type among the plurality of predetermined document types the identification target document corresponds to (in step 1725, the document is subsequently classified based on the comparison [col. 20 | ln. 51-56; Fig. 17]).
Regarding claim 13, Sampson et al teach The information processing system of claim 1 (as previously described), wherein
the predetermined document type is one of a plurality of predetermined document types ((a set of document classes 430 and associated document template 435 is output by training module 425 [col. 4 | ln. 50-54; Fig. 4]), and
the circuitry (central processor 302 [col. 3 | ln. 53-55; Fig. 3]) is configured to:
store a frequently occurring word string of each of the plurality of predetermined document types (in step 1710, the system creates and stores templates corresponding to each document class, with each template including a set of keywords [col. 20 | ln. 37-43; Fig. 17]);
acquire information on a position of the frequently occurring word string of each of the plurality of predetermined document types in the identification target document (in step 1120, the location of a word, such as a keyword, is identified in a first document [col. 13 | ln. 46-54; Fig. 11]);
generate a feature quantity of the identification target document using the information on the position (in step 1120, the relative position of a word, such as a keyword, is identified relative to other words in the document [col. 13 | ln. 46-54; Fig. 11]), the feature quantity including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string of each of the plurality of predetermined document types and another word string in the identification target document (in step 1120, the relative position of a word, such as a keyword, is identified relative to other words in the document [col. 13 | ln. 46-54; Fig. 11]);
store a trained model that identifies the plurality of predetermined document types (system 405 comprises a training module 410, a classification module 415, and a word location comparison engine 420, which is used to group and classify documents [col. 4 | ln. 50-54; Fig. 4]),
the trained model being generated through machine learning (the system 405 includes a training module 415 that receives a set of training documents 425 to output a set of document classes 430 and associated document templates 435, which can be generated via automatic training [col. 4 | ln. 52-67 & col. 5 | ln. 1 – 15; Fig. 4], and that during the automatic training step, a distance function is used to determine whether a document is “close” to a template [col. 5 | ln. 62 – 65], which is further outlined in Fig. 9, with step 920 outlining a clustering algorithm to compare and group similar documents [col. 8 | ln. 1-12]) such that, in response to input of a feature quantity of a document including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string of each of the plurality of predetermined document types and another word string in the document (a comparison function takes two images and finds a set of common words in approximately the same relative position in each document [col. 8 | ln. 23-30], with step 1720 comparing the position of keywords with other words in the document [col. 20 | ln. 49-54]), information indicating appropriateness of the document being a document of each of the plurality of predetermined document types is output (a scoring function provides a score proportional to a number and size of the common word across each document [col. 8 | ln. 29-33]); and
input the feature quantity of the identification target document to the trained model that identifies the plurality of predetermined document types, to identify which document type among the plurality of predetermined document types the identification target document corresponds to (in step 1725, the spatial relationship between a keyword and another word is provided to the training model to compare to a template and classify the document according to that template [col. 20 | ln. 49-54; Fig. 17]).
Regarding claim 15, Sampson et al teach The information processing system of claim 13 (as described above), wherein
the positional relationship feature quantity includes a feature quantity indicating a distance between frequently occurring word strings of a combination satisfying a predetermined condition among combinations of two frequently occurring word strings of the predetermined document type (a combination of words and their positional relationships (represented as distance vectors) across differing documents is compared [col. 14 | ln. 29-62; Fig. 14]), and
the combination satisfying the predetermined condition is a combination of frequently occurring word strings for which a representative value of a distance between the frequently occurring word strings in the plurality of training images that are images of the predetermined document type is less than or equal to a certain value (the system then generates a list of all word distance comparisons which is scored and then compared to a threshold value or score to determine if the documents are of a matching class [col. 14 | ln. 43-62]).
Regarding claim 16, Sampson et al disclose a system and method for classifying documents using optical character recognition (OCR) data. More specifically, Sampson et al teach An information processing system (system 405 for grouping and classifying documents [col. 4 | ln. 50-53; Fig. 4]) comprising:
circuitry (central processor 302 [col. 3 | ln. 53-55; Fig. 3]) configured to:
acquire a character recognition result (receive OCR data of a document image [col. 7 | ln. 8-13]) of each of a plurality of training images including a plurality of predetermined document type images that are images of documents of a predetermined document type having layouts different from one another (a class of documents is derived from a set of training documents of a known predetermined type in step 905 [col. 6 | ln. 49-51; Fig. 9], with Figs. 6 & 7 illustrating two documents of the same type (in this case, an invoice) with different layouts, but the same keywords [col. 6 | ln. 21-48]);
acquire a frequently occurring word string of the predetermined document type (in step 1710, the system creates and stores templates corresponding to each document class, with each template including a set of keywords [col. 20 | ln. 37-43; Fig. 17]);
detect the frequently occurring word string from the character recognition result of each of the plurality of training images to acquire information on a position of the frequently occurring word string in a document depicted in the training image (in steps 1715 & 1720, the system receives an input document for classification and compares the spatial relationship of keywords to other words in the document based on their location in the document [col. 20 | ln. 44-56; Fig. 17]);
generate a feature quantity of the document depicted in the training image using the information on the position of the frequently occurring word string in the document depicted in each of the plurality of training images (in step 1720, the system obtains information regarding the location of a keyword and different words in a template derived from training images [[col. 20 | ln. 44-56; Fig. 17]), the feature quantity including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string and another word string in the document depicted in the training image (in step 1720, the system compares the relationships of a keyword and a different word in the template [col. 20 | ln. 49-54]); and
generate a trained model that identifies the predetermined document type (system 405 comprises a training module 410, a classification module 415, and a word location comparison engine 420, which is used to group and classify documents [col. 4 | ln. 50-54; Fig. 4]), the trained model being generated through machine learning using training data (the system 405 includes a training module 415 that receives a set of training documents 425 to output a set of document classes 430 and associated document templates 435, which can be generated via automatic training [col. 4 | ln. 52-67 & col. 5 | ln. 1 – 15; Fig. 4], and that during the automatic training step, a distance function is used to determine whether a document is “close” to a template [col. 5 | ln. 62 – 65], which is further outlined in Fig. 9, with step 920 outlining a clustering algorithm to compare and group similar documents [col. 8 | ln. 1-12]),
the training data associating the feature quantity of the document depicted in each of the plurality of training images with information indicating whether the document depicted in the training image is a document of the predetermined document type (using the spatial relationship compared in step 1720, the document is classified in step 1725 [col. 20 | ln. 51-56; Fig. 17]).
Regarding claim 17, Sampson et al teach The information processing system of claim 16 (as described above), wherein the circuitry (central processor 302 [col. 3 | ln. 53-55; Fig. 3]) is configured to:
extract a word string that appears in documents depicted in the plurality of predetermined document type images (an algorithm extracts common words and adds it to a list [col. 13 | ln. 57-67]), based on the character recognition results of the plurality of predetermined document type images (the comparison function takes OCR data that may include a set of characters with their position and confidence information from two images [col. 8 | ln. 22-35]); and
acquire the extracted word string as the frequently occurring word string of the predetermined document type (scoring of the common words based on their location and size is used to determine the document class [col. 8 | ln. 22-35]).
Regarding claim 18, Sampson et al teach The information processing system of claim 16 (as previously described), wherein the circuitry (central processor 302 [col. 3 | ln. 53-55; Fig. 3]) is configured to:
acquire a ground truth definition in which identification information of each of the plurality of training images is associated with information indicating whether a document depicted in the training image is a document of the predetermined document type (during training, either automated or manual, a set of documents with known classes are used to generate document classes and templates [col. 5 | ln. 11-20]); and
acquire, based on the ground truth definition, the information indicating whether a document depicted in a training image among the plurality of training images is a document of the predetermined document type (during training, either automated or manual, a set of documents with known classes are used to generate document classes and templates [col. 5 | ln. 11-20]).
Regarding claim 19, Sampson et al disclose a system and method for classifying documents using optical character recognition (OCR) data. More specifically, Sampson et al teach A document type identification method (document classification method flow outlined in Fig. 17 [col. 20 | ln. 37-56]) comprising:
acquiring a character recognition result of an identification target image that is an image of an identification target document (receive OCR data of a document image [col. 7 | ln. 8-13]);
storing a frequently occurring word string of a predetermined document type (in step 1710, the system creates and stores templates corresponding to each document class, with each template including a set of keywords [col. 20 | ln. 37-43; Fig. 17]);
detecting the frequently occurring word string from the character recognition result of the identification target image to acquire information on a position of the frequently occurring word string in the identification target document (in steps 1715 & 1720, the system receives an input document for classification and compares the spatial relationship of keywords to other words in the document based on their location in the document [col. 20 | ln. 44-56; Fig. 17]);
generating a feature quantity of the identification target document using the information on the position, the feature quantity including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string and another word string in the identification target document (the system obtains information regarding the location of both a keyword and a different word in the document to be compared, and compares it to their relationships in the template document [col. 20 | ln. 49-54]; this is further exemplified in Fig. 14, illustrating the spatial relationship of words across two documents [col. 14 | ln. 1-9]);
storing a trained model that identifies the predetermined document type (system 405 comprises a training module 410, stored in the system memory 304 [col. 3 | ln. 45-55 & col. 4 | ln. 50-54; Figs. 3 & 4]), the trained model being generated through machine learning (the system 405 includes a training module 415 that receives a set of training documents 425 to output a set of document classes 430 and associated document templates 435, which can be generated via automatic training [col. 4 | ln. 52-67 & col. 5 | ln. 1 – 15; Fig. 4], and that during the automatic training step, a distance function is used to determine whether a document is “close” to a template [col. 5 | ln. 62 – 65], which is further outlined in Fig. 9, with step 920 outlining a clustering algorithm to compare and group similar documents [col. 8 | ln. 1-12]) such that, in response to input of a feature quantity of a document including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string and another word string in the document (a comparison function takes two images and finds a set of common words in approximately the same relative position in each document [col. 8 | ln. 23-30], with step 1720 comparing the position of keywords with other words in the document [col. 20 | ln. 49-54]), information indicating appropriateness of the document being a document of the predetermined document type is output (a scoring function provides a score proportional to a number and size of the common word across each document [col. 8 | ln. 29-33]); and
inputting the feature quantity of the identification target document to the trained model to identify whether the identification target document is a document of the predetermined document type (in step 1725, the spatial relationship between a keyword and another word is provided to the training model to compare to a template and classify the document according to that template [col. 20 | ln. 49-54; Fig. 17]).
Regarding claim 20, Sampson et al disclose a system and method for classifying documents using optical character recognition (OCR) data. More specifically, Sampson et al teach A model generation method (model training method outlined in Fig. 9 [col 6 | ln. 49-61]) comprising:
acquiring a character recognition result (receive OCR data of a document image [col. 7 | ln. 8-13]) of each of a plurality of training images including a plurality of predetermined document type images that are images of documents of a predetermined document type having layouts different from one another (a class of documents is derived from a set of training documents of a known predetermined type in step 905 [col. 6 | ln. 49-51; Fig. 9], with Figs. 6 & 7 illustrating two documents of the same type (in this case, an invoice) with different layouts, but the same keywords [col. 6 | ln. 21-48]);
acquiring a frequently occurring word string of the predetermined document type (in step 1710, the system creates and stores templates corresponding to each document class, with each template including a set of keywords [col. 20 | ln. 37-43; Fig. 17]);
detecting the frequently occurring word string from the character recognition result of each of the plurality of training images to acquire information on a position of the frequently occurring word string in a document depicted in the training image (in steps 1715 & 1720, the system receives an input document for classification and compares the spatial relationship of keywords to other words in the document based on their location in the document [col. 20 | ln. 44-56; Fig. 17]);
generating a feature quantity of the document depicted in the training image, using the information on the position of the frequently occurring word string in the document depicted in each of the plurality of training images (in step 1720, the system obtains information regarding the location of a keyword and different words in a template derived from training images [[col. 20 | ln. 44-56; Fig. 17]), the feature quantity including a positional relationship feature quantity related to a positional relationship between the frequently occurring word string and another word string in the document depicted in the training image (in step 1720, the system compares the relationships of a keyword and a different word in the template [col. 20 | ln. 49-54]); and
generating a trained model that identifies the predetermined document type (system 405 comprises a training module 410, a classification module 415, and a word location comparison engine 420, which is used to group and classify documents [col. 4 | ln. 50-54; Fig. 4]), the trained model being generated through machine learning using training data (the system 405 includes a training module 415 that receives a set of training documents 425 to output a set of document classes 430 and associated document templates 435, which can be generated via automatic training [col. 4 | ln. 52-67 & col. 5 | ln. 1 – 15; Fig. 4], and that during the automatic training step, a distance function is used to determine whether a document is “close” to a template [col. 5 | ln. 62 – 65], which is further outlined in Fig. 9, with step 920 outlining a clustering algorithm to compare and group similar documents [col. 8 | ln. 1-12]),
the training data associating the feature quantity of the document depicted in each of the plurality of training images with information indicating whether the document depicted in the training image is a document of the predetermined document type (using the spatial relationship compared in step 1720, the document is classified in step 1725 [col. 20 | ln. 51-56; Fig. 17]).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 9 & 10 are rejected under 35 U.S.C. § 103 as being unpatentable over Sampson et al (US 8724907 B1) in view of Moneer; Alitto (EP 3428844 A1).
Regarding claim 9, Sampson et al teach The information processing system of claim 8 (as described previously), wherein the circuitry (Sampson et al: central processor 302 [col. 3 | ln. 53-55; Fig. 3]) is configured to:
in a case where the identification target document is identified to be a document of two or more predetermined document types as a result of identification performed for each of the plurality of predetermined document types (Sampson et al: if the document belongs to more than one class, it may be considered over-classified [col. 5 | ln. 54-61]), but do not teach selecting a document type when two or more document types are identified.
Moneer; Alitto, however, is analogous art pertinent to the field of the endeavor of the present application and disclose a method for selecting a document type when multiple document types are identified. More specifically, Moneer; Alitto teaches
select a document type from the two or more predetermined document types (Moneer; Alitto: determining the document type of the best class as the document type [¶001600]); and
determine the selected document type as the document type of the identification target document (Moneer; Alitto: a probability of the signature (a feature quantity used for comparison) is calculated for the document being in a plurality of document type classes, and identifying the best document type class based on the probabilities [¶0015-16]). Thus, in accordance with KSR rationales (see MPEP § 2143), the prior art includes all of the claimed elements in the present application, with the only difference being the lack of combination. Furthermore, one of ordinary skill in the art could have easily combined the elements by known methods and that each element would merely perform the same function as it does separately. For example, the inclusion of further selecting a document type when two or more document types are applicable enables the invention to still output a document type, improving overall operability. Furthermore, selection of a document class based on the probability a document belongs in a given class is well known in the art. One of ordinary skill in the art would recognize that the result of the combination would be predictable, since the added selection criteria would merely improve document classification efficiency, resulting in an improved process.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to take the document identification method proposed by Sampson et al and implement the selection of a document type based on probability by Moneer; Alitto to arrive at the invention of the present application.
Regarding claim 10, Sampson et al, in view of Moneer; Alitto teach The information processing system of claim 9 (as described above), wherein the circuitry (Sampson et al: central processor 302 [col. 3 | ln. 53-55; Fig. 3]) is configured to select a document type from the two or more predetermined document types (Moneer; Alitto: determining the document type of the best class as the document type [¶0015]), based on a probability of the identification target document being a document of each of the two or more predetermined document type (Moneer; Alitto: a probability of the signature (a feature quantity used for comparison) is calculated for the document being in a plurality of document type classes, and identifying the best document type class based on the probabilities [¶0015]). Thus, in accordance with KSR rationales (see MPEP § 2143), the prior art includes all of the claimed elements in the present application, with the only difference being the lack of combination. Furthermore, one of ordinary skill in the art could have easily combined the elements by known methods and that each element would merely perform the same function as it does separately. For example, the inclusion of further selecting a document type when two or more document types are applicable enables the invention to still output a document type, improving overall operability. Furthermore, selection of a document class based on the probability a document belongs in a given class is well known in the art. One of ordinary skill in the art would recognize that the result of the combination would be predictable, since the added selection criteria would merely improve document classification efficiency, resulting in an improved process.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to take the document identification method proposed by Sampson et al and implement the selection of a document type based on probability by Moneer; Alitto to arrive at the invention of the present application.
Claim 14 is rejected under 35 U.S.C. § 103 as being unpatentable over Sampson et al (US 8724907 B1) in view of Fusco et al (US 2022/0075809).
Regarding claim 14, Sampson et al teach The information processing system of claim 13 (as described above), the positional relationship feature quantity is a positional relationship feature quantity related to a positional relationship between the frequently occurring word string of each of the plurality of document types and another word string (Sampson et al: in step 1120, the relative position of a word, such as a keyword, is identified relative to other words in the document [col. 13 | ln. 46-54; Fig. 11]), but does not teach handling for when the frequently occurring word string occurs across multiple document types.
Fusco et al, on the other hand, is analogous art pertinent to the field of endeavor of the present application and disclose a bootstrap document classifier that ignores matching words that appear across multiple keyword sets for different document types. More specifically, Fusco et al teach wherein in a case where the plurality of predetermined document types have an overlapping frequently occurring word string (Fusco et al: in step 58, the search module 31 checks whether any keywords from step 55 were identified across more than one keyword set (the keyword set corresponding to keywords for an associated document type) [¶0048; Fig. 4]), the frequently occurring word string of each of the predetermined document types being not the overlapping frequently occurring word string (Fusco et al: any keywords identified across more than one keyword set are excluded [¶0048; step 58 of Fig. 4]). Furthermore, Fusco et al disclose that this eliminated keywords which are potentially non-discriminative, which improves the overall quality of the resulting dataset [¶0010].
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to implement the keyword exclusion criteria outlined by Fusco et al to improve the class identification system to outlined by Sampson et al to arrive at the invention of the present application.
Claim 11 is rejected under 35 U.S.C. § 103 as being unpatentable over Sampson et al (US 8724907 B1) in view of Moneer; Alitto (EP 3428844 A1), further in view of Bianchi et al (US 2021/0303627 A1).
Regarding claim 11, Sampson et al teach The information processing system of claim 9 (as previously described), wherein the circuitry (Sampson et al: central processor 302 [col. 3 | ln. 53-55; Fig. 3]) but does not teach selecting the document type from two or more predetermined types.
Moneer; Alitto, however, is analogous art pertinent to the field of the endeavor of the present application and disclose a method for selecting a document type when multiple document types are identified. More specifically, Moneer; Alitto teaches is configured to select a document type from the two or more predetermined document types, (Moneer; Alitto: determining the document type of the best class as the document type [¶0015]). Thus, in accordance with KSR rationales (see MPEP § 2143), the prior art includes all of the claimed elements in the present application, with the only difference being the lack of combination. Furthermore, one of ordinary skill in the art could have easily combined the elements by known methods and that each element would merely perform the same function as it does separately. For example, the inclusion of further selecting a document type when two or more document types are applicable enables the invention to still output a document type, improving overall operability. Furthermore, selection of a document class based on the probability a document belongs in a given class is well known in the art. One of ordinary skill in the art would recognize that the result of the combination would be predictable, since the added selection criteria would merely improve document classification efficiency, resulting in an improved process.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to take the document identification method proposed by Sampson et al and implement the selection of a document type based on probability by Moneer; Alitto to arrive at the invention of the present application.
Additionally, Sampson et al in view of Moneer; Alitto does not teach selection based on a number of times the document was identified previously. Bianchi et al, on the other hand, is analogous art pertinent to the field of endeavor of the present invention and disclose a text and document classifier that takes into account previous predetermined document types. More specifically, Bianchi et al teach based on a number of times each of the two or more predetermined document types was identified as the document type of the identification target document by the trained model in past (Bianchi et al: in step 650 of the document classifier workflow, the frequency of previously identified document classes are weighted when identifying a current document [¶0049 & 52-53; Fig. 6]). Additionally, Bianchi et al disclose that their approach minimizes retraining and improves overall accuracy through their Bayesian classifier [¶0035].
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to take the document identification and type selection system proposed by Sampson et al in view of Moneer; Alitto to modify it with the classification step factoring in the frequency of previous document types outlined by Bianchi et al to arrive at the invention of the present application.
Claim 12 is rejected under 35 U.S.C. § 103 as being unpatentable over Sampson et al (US 8724907 B1) in view of Moneer; Alitto (EP 3428844 A1), further in view of Freed et al (US 2020/0125827 A1).
Regarding claim 12, Sampson et al teach The information processing system of claim 9 (as previously described), wherein the circuitry (Sampson et al: central processor 302 [col. 3 | ln. 53-55; Fig. 3]) but does not teach selecting the document type from two or more predetermined types.
Moneer; Alitto, however, is analogous art pertinent to the field of the endeavor of the present application and disclose a method for selecting a document type when multiple document types are identified. More specifically, Moneer; Alitto teaches is configured to select a document type from the two or more predetermined document types, (Moneer; Alitto: determining the document type of the best class as the document type [¶0015]). Thus, in accordance with KSR rationales (see MPEP § 2143), the prior art includes all of the claimed elements in the present application, with the only difference being the lack of combination. Furthermore, one of ordinary skill in the art could have easily combined the elements by known methods and that each element would merely perform the same function as it does separately. For example, the inclusion of further selecting a document type when two or more document types are applicable enables the invention to still output a document type, improving overall operability. Furthermore, selection of a document class based on the probability a document belongs in a given class is well known in the art. One of ordinary skill in the art would recognize that the result of the combination would be predictable, since the added selection criteria would merely improve document classification efficiency, resulting in an improved process.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to take the document identification method proposed by Sampson et al and implement the selection of a document type based on probability by Moneer; Alitto to arrive at the invention of the present application.
Additionaly, Sampson et al in view of Moneer; Alitto do not teach selection based on a number of times the document was identified previously. Freed et al, however, is analogous art pertinent to the field of endeavor of the present invention and disclose a classifier that takes into account a sequential probability for when a document is recorded. More specifically, Freed et al teach based on a timing at which each of the two or more predetermined document types was identified as the document type of the identification target document by the trained model (Freed et al: the classifier 130 of Fig. 1 performs classification based on a sequential probability of the expected time windows the document is recorded derived from an extracted date [¶0032]). Furthermore, Freed et al disclose that accounting for a date a document is recorded provides a “context” for classification to increase accuracy [¶0032], with this context being of particular importance when handling documents that may be received at predictable intervals over a set period of time, and/or when many types of documents of the same type recorded sequentially over a short period of time [¶0021-23].
Therefore, it would have been obvious before the effective filing date of the present application to incorporate the teachings of Freed et al to provide a “timing-aware” selection procedure to improve the document classifier taught by Sampson et al in view of Moneer; Alitto to arrive at the invention of the instant application.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Tatsumi et al (US 2022/0180091 A1) disclose a system for identifying key character strings in a target image of a document.
Zhang et al (US 2020/0311413 A1) disclose a system for identifying form types utilizing a set of keywords.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Michael M. Sofroniou whose telephone number is (571)272-0287. The examiner can normally be reached M-F: 8:30 AM - 5:00 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, John M. Villecco can be reached at (571) 272-7319. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MICHAEL M SOFRONIOU/Examiner, Art Unit 2661
/JOHN VILLECCO/Supervisory Patent Examiner, Art Unit 2661