Categorizer assisted capture of customer documents using a mobile device

Patent metadata

Field	Value
Publication number	US-2016104040-A1
Application number	US-201414509357-A
Country	US
Kind code	A1
Filing date	Oct 8, 2014
Priority date	Oct 8, 2014
Publication date	Apr 14, 2016
Grant date	—

Abstract

A computer-implemented system and method for controlling document image capture are provided. The method includes identifying a categorization model for categorizing image frames and configuring an image capture device, based on the identified categorization model. A flow of frames of a same document captured with the configured image capture device is received and an image representation generated for each of a plurality of frames within the flow of frames. With the identified categorization model, for each of the plurality of frames, a confidence for at least one category is assigned to the frame based on the image representation. A category is assigned to the document when a threshold confidence for one of the at least one categories is assigned. An automatic capture of an image of the document is triggered based on the assigned category.

First claim

What is claimed is:

1 . A method for controlling document image capture comprising:
identifying a categorization model for categorizing image frames; configuring an image capture device, based on the identified categorization model; receiving a flow of frames of a same document captured with the configured image capture device; generating an image representation for each of a plurality of frames within the flow of frames; with the identified categorization model, for each of the plurality of frames, assigning a confidence for at least one category to the frame based on the image representation; and assigning a category to the document when a threshold confidence for one of the at least one categories is assigned; and triggering an automatic capture of an image of the document based on the assigned category, wherein at least one of the configuring of the image capture device, generating of the image representation, assigning a confidence, and assigning a category is performed by a processor.

2 . The method of claim 1 , wherein the identifying of the categorization model comprises providing for a user to load a new categorization model.

3 . The method of claim 1 , further comprising storing a set of parameters for the categorization model, the parameters comprising at least one of document dimensions, quality, and form layout, the parameters being compared with the document frames in the configuring of the image capture device.

4 . The method of claim 1 , further comprising adjusting the image capture device in real-time in response to external conditions, based on the identified categorization model.

5 . The method of claim 1 , further comprising training the categorization model on a set of training images, the set of training images each comprising a category label and an image signature, the labels being selected from a finite set of image categories.

6 . The method of claim 1 , further comprising providing for forcing the capture of a frame via the image capture device when the score for at least one of the plurality of frames does not meet the threshold.

7 . The method of claim 6 , further comprising providing for a user to manually assign a category to the forced captured frame.

8 . The method of claim 1 , wherein the image representations are based on runlength histograms.

9 . The method of claim 1 wherein the category is assigned only when each of a plurality of frames meets the threshold confidence for that category.

10 . The method of claim 1 , wherein the document includes a set of documents and the method includes for each document in turn, configuring the image capture device based on one of the categories, receiving a flow of frames, generating image representations, assigning a confidence, and triggering the automatic capture of the respective document.

11 . The method of claim 1 , wherein the configuring of the image capture device, generating of the image representation, assigning a confidence, and assigning a category are performed on a mobile device which includes the image capture device.

12 . The method of claim 1 , wherein the at least one category comprises a plurality of categories.

13 . The method of claim 1 , further comprising:
storing a set of sample images, each of the sample images in the set corresponding to a respective one of each of the categories; and displaying a semi-transparent overlay of one of the sample images in the set for assisting a user in positioning the image capture device for capturing the image.

14 . The method of claim 13 , further comprising:
validating the assigned category with an image representation of the displayed semi-transparent overlay of the sample image.

15 . The method of claim 1 , further comprising:
sending the captured image to a server for processing after validating that the captured document image matches one of the at least one categories.

16 . A computer program product comprising a non-transitory recording medium which stores instructions which when executed by a computer, perform the method of claim 1 .

17 . A system comprising memory which stores instructions for performing the method of claim 1 and a processor in communication with the memory for executing the instructions.

18 . A system for controlling document image capture comprising:
an image representation generator which receives a flow of frames of a same document captured with an image capture device and generates an image representation for each of a plurality of frames within the flow of frames; a categorizer for categorizing image frames with a categorization model, for each of the plurality of frames, the categorizer assigning a confidence for at least one category to the frame based on the respective image representation and the model, the categorizer assigning a category to the document when a threshold confidence for one of the at least one category is assigned; a capture control component for configuring the image capture device, the capture control component triggering an automatic capture of an image of the document with the image capture device based on the assigned category; and a processor which implements the image representation generator, categorizer, and capture control component.

19 . The system of claim 18 , wherein the capture control component configures the image capture device based on parameters stored in the categorization model.

20 . A method comprising:
configuring an image capture device of a mobile device, based on parameters of a categorization model; receiving a flow of frames of a same document captured with the configured image capture device; with a representation generator of the mobile device, generating an image representation for each of a plurality of frames within the flow of frames; with a categorizer of the mobile device, for each of the plurality of frames, assigning a confidence for at least one category to the frame based on the image representation; and assigning a category to the document when a threshold confidence for one of the at least one categories is assigned to a plurality of frames; and triggering automatic capture of an image of the document based on the assigned category, wherein at least one of the configuring of the image capture device, generating of the image representation, assigning a confidence, and assigning a category is performed by a processor.

Assignees

• Xerox Corp

Inventors

• Ragnet François
• Cramet Damien

Classifications

• G06V30/40 Primary

  Document-oriented image-based pattern recognition · CPC title

• H04N1/107

  with manual scanning · CPC title

• G06V10/235

  based on user input or interaction · CPC title

• H04N1/00249

  with a photographic apparatus, e.g. a photographic printer or a projector (photographic apparatus per se G03B, G03D) · CPC title

• H04N1/40062

  Discrimination between different image types, e.g. two-tone, continuous tone · CPC title