Multi-modal document type classification systems and methods
Artificial intelligence (AI) systems of the inventive subject matter are directed to receiving uploaded images containing one or more document, identifying the type of documents received, and returning information contained in those documents. Upon receiving an image containing a document, the AI system: checks for barcodes, OCRs any text, detects visual features, and detects an overall document shape. The AI system can then use any information gathered during those steps to ultimately verify information contained in the document.
1 . A method of using an artificial intelligence (AI) system to classify documents, the method comprising the steps of:
receiving, at a platform server running the AI system, a file comprising a document;
receiving, at the platform server, a user-specified document type, wherein the user-specified document type comprises a set of potential document types;
identifying, using the user-specified document type, a set of document candidates;
conducting optical character recognition (OCR) on the document to extract text content of the document;
identifying, based on the text content, a document candidate from the set of document candidates, wherein the document candidate has an expected document shape;
classifying the document as the document candidate at least in part by determining that an overall shape of the document matches the expected document shape.
2 . The method of claim 1 , wherein the step of identifying the document candidate based on the text content relies on the AI system having been trained using a training set of known document types.
3 . The method of claim 1 , further comprising the step of detecting, by the AI system, a first filetype and, based on the first filetype, determining whether to convert the file to a second filetype.
4 . The method of claim 1 , further comprising the step of detecting and decoding a machine-readable zone.
5 . A method of using an artificial intelligence (AI) system to classify documents, the method comprising the steps of:
receiving, at a platform server running the AI system, a file comprising a document;
receiving, at the platform server, a user-specified document type, wherein the user-specified document type indicates a document species;
wherein the document species is associated with expected content categories;
classifying the document as the document species; and
extracting data from the document, wherein portions of the data is selectively matched to the expected content categories.
6 . The method of claim 5 , where the document species is a single type of document.
7 . The method of claim 5 , further comprising the step of detecting, by the AI system, a first filetype and, based on the first filetype, determining whether to convert the file to a second filetype.
8 . A method of using an artificial intelligence (AI) system to classify documents, the method comprising the steps of:
receiving, at a platform server running the AI system, a file comprising a document;
receiving, at the platform server, a user-specified document type, wherein the user-specified document type comprises a set of potential document types;
visually identifying that a barcode is present on the document;
decoding the barcode to extract barcode data;
identifying, using the barcode data, a document candidate from the set of document candidates, wherein the document candidate has an expected document shape; and
classifying the document as the document candidate at least in part by determining that an overall shape of the document matches the expected document shape.
9 . The method of claim 8 , wherein the barcode is a matrix type barcode.
10 . The method of claim 8 , wherein the barcode data comprises text information that the AI system has been trained to associate with the document candidate.
11 . The method of claim 8 , further comprising the step of determining document validity using the barcode data.
12 . The method of claim 8 , wherein the step of identifying the document candidate relies on the AI system being trained to detect the document candidate using a training set of known documents that includes the document candidate.