Hachage d'image perceptuel partiel pour la déconstruction de factures.

L'invention a pour objet un appareil et un procédé pour identifier le fournisseur dans une facture (101) en divisant la facture en trois régions (102-104) et en effectuant un hachage d'image perceptuel sur chaque section. Ensuite, une distance de Hamming est utilisée pour comparer la valeur de hachage de chaque section avec les valeurs de hachages de factures connues afin d'identifier le fournisseur qui a envoyé la facture.

Partial perceptual image hashing for invoice deconstruction

Partial perceptual image hashing for invoice deconstruction

An image of an invoice (e.g. rasterised document, PNG image etc, may contain logo in the header) is split into regions (302, see top, middle and bottom parts 102, 103 and 104 of the invoice, fig 1). A perceptual image hash is calculated from each region of the invoice image (303). Each hash is then matched against an entry in a database where the vendor name is known (look up match 304) by calculating a hamming distance between an image hash and each entry in the database of known vendors and taking the smallest hamming distance. In this way, the database is searched for a match and the vendor associated with the invoice is identified. The perceptual image hash may be calculated using: an average or difference or a pHash algorithm. The invoice image may be reduced to greyscale and to an 8x8 grid of pixels. If the sum of the best two hamming distances of the hashes of two of the invoice regions is greater than a threshold, then the invoice 101 is not considered to match the database of known ...

Partial perceptual image hashing for invoice deconstruction

A system and method for deconstructing a document is described herein, where the method is an improvement over existing document deconstruction techniques. These improvements increase speed and accuracy by rapidly identifying the vendor in an invoice by splitting the invoice into three regions and performing a perceptual image hashing on each section. Then a hamming distance is used to compare the hash for each section with the hashes of known invoices to identify the vendor who sent the invoice.

Partial Perceptual Image Hashing for Document Deconstruction

A system and method for deconstructing a document is described herein, where the method is an improvement over existing document deconstruction techniques. These improvements increase speed and accuracy by rapidly identifying the source in a document by splitting the document into a plurality of sections and performing a perceptual image hashing on each section. Then a hamming distance is used to compare the hash for each section with the hashes of known documents to identify the source who sent the document.

Partial Perceptual Image Hashing for Invoice Deconstruction

A system and method for deconstructing a document is described herein, where the method is an improvement over existing document deconstruction techniques. These improvements increase speed and accuracy by rapidly identifying the vendor in an invoice by splitting the invoice into three regions and performing a perceptual image hashing on each section. Then a hamming distance is used to compare the hash for each section with the hashes of known invoices to identify the vendor who sent the invoice.