Back to results
US 11,961,317 B2Granted
Extracting textual information from image documents
AssigneeOracle Financial Services Software Limited
InventorsDakshayani Singaraju, Veresh Jain, Kartik Kumar
Life of this patent
Estimating the term…
Abstract
Aspects of the present disclosure are directed to extracting textual information from image documents. In one embodiment, a system, upon receiving a request to extract textual information from an image document, a digital processing system performs character recognition based on content of the image document using multiple approaches to generate corresponding texts. The texts are then combined to determine a result text representing the textual information contained in the image document. The result is then provided as a response to the request.
Claims20
A non-transitory machine-readable medium storing one or more sequences of instructions for extracting textual information from image documents, wherein execution of said one or more instructions by one or more processors contained in a digital processing system cause said digital processing system to perform the actions of:
- receiving a request to extract textual information from an image document containing a plurality of fields, each field having a corresponding image portion containing a key portion and a value portion, said image document being encoded according to a set of image document properties;
- identifying, at a first time instance, a plurality of image enhancements from a list of image enhancements based on said set of image document properties, wherein said set of image document properties comprises at least one of source machine type generating said image document, a resolution of said image document and a number and types of colors in said image document,
- wherein each image enhancement comprises a corresponding sequence of image processing techniques, wherein each sequence of image processing techniques comprises two or more of Denoising, Blurring, Grayscaling, Sharpening, Binarization, Contrasting, Enhancing, Brightening, Shadow Removal, Dilation, Erosion and Blurring;
- performing, after said first time instance, character recognition based on images of said image document using a plurality of approaches executed independently, wherein each approach of said plurality of approaches comprises applying a corresponding image enhancement of said plurality of image enhancements, wherein applying an image enhancement comprises applying in series the corresponding sequence of image processing techniques of the image enhancement,
- wherein character recognition using each approach generates a corresponding recognized text for each field, the recognized text containing a key and a value respectively representing text in the key portion and the value portion of the field, each value comprising a plurality of characters,
- wherein a respective plurality of values are generated for each field of the plurality of fields, wherein each value of said respective plurality of values is generated for the field by said performing character recognition using a corresponding approach of said plurality of approaches;
- for each field of said plurality of fields, selecting one value of said respective plurality of values for the field as representing a recognized best value for the field, wherein said selecting selects a first value generated by a first sequence of image processing techniques as the recognized best value for a first field and a second value generated by a second sequence of image processing techniques as the recognized best value for a second field, wherein said first field and said field are comprised in said plurality of fields;
- combining the respective recognized best value and said key for each field of said plurality of fields to form a result text representing the textual information contained in said image document;
- and
- providing said result text as a response to said request.
7 more elements
A computer implemented method for extracting textual information from image documents, said method comprising:
- receiving a request to extract textual information from an image document containing a plurality of fields, each field having a corresponding image portion containing a key portion and a value portion, said image document being encoded according to a set of image document properties;
- identifying, at a first time instance, a plurality of image enhancements from a list of image enhancements based on said set of image document properties,
- wherein said set of image document properties comprises at least one of source machine type generating said image document, a resolution of said image document and a number and types of colors in said image document,
- wherein each image enhancement comprises a corresponding sequence of image processing techniques, wherein each sequence of image processing techniques comprises two or more of Denoising, Blurring, Grayscaling, Sharpening, Binarization, Contrasting, Enhancing, Brightening, Shadow Removal, Dilation, Erosion and Blurring;
- performing, after said first time instance, character recognition based on images of said image document using a plurality of approaches executed independently, wherein each approach of said plurality of approaches comprises applying a corresponding image enhancement of said plurality of image enhancements, wherein applying an image enhancement comprises applying in series the corresponding sequence of image processing techniques of the image enhancement,
- wherein character recognition using each approach generates a corresponding recognized text for each field, the recognized text containing a key and a value respectively representing text in the key portion and the value portion of the field, each value comprising a plurality of characters,
- wherein a respective plurality of values are generated for each field of the plurality of fields, wherein each value of said respective plurality of values is generated for the field by said performing character recognition using a corresponding approach of said plurality of approaches;
- for each field of said plurality of fields, selecting one value of said respective plurality of values for the field as representing a recognized best value for the field, wherein said selecting selects a first value generated by a first sequence of image processing techniques as the recognized best value for a first field and a second value generated by a second sequence of image processing techniques as the recognized best value for a second field, wherein said first field and said field are comprised in said plurality of fields;
- combining the respective recognized best value and said key for each field of said plurality of fields to form a result text representing the textual information contained in said image document;
- and
- providing said result text as a response to said request.
8 more elements
A digital processing system comprising:
- a random access memory (RANI) to store instructions for extracting textual information from image documents;
- and
- one or more processors to retrieve and execute the instructions, wherein execution of the instructions causes the digital processing system to perform the actions of:
- receiving a request to extract textual information from an image document containing a plurality of fields, each field having a corresponding image portion containing a key portion and a value portion, said image document being encoded according to a set of image document properties;
- identifying, at a first time instance, a plurality of image enhancements from a list of image enhancements based on said set of image document properties,
- wherein said set of image document properties comprises at least one of source machine type generating said image document, a resolution of said image document and a number and types of colors in said image document,
- wherein each image enhancement comprises a corresponding sequence of image processing techniques, wherein each sequence of image processing techniques comprises two or more of Denoising, Blurring, Grayscaling, Sharpening, Binarization, Contrasting, Enhancing, Brightening, Shadow Removal, Dilation, Erosion and Blurring;
- performing, after said first time instance, character recognition based on images of said image document using a plurality of approaches executed independently, wherein each approach of said plurality of approaches comprises applying a corresponding image enhancement of said plurality of image enhancements, wherein applying an image enhancement comprises applying in series the corresponding sequence of image processing techniques of the image enhancement,
- wherein character recognition using each approach generates a corresponding recognized text for each field, the recognized text containing a key and a value respectively representing text in the key portion and the value portion of the field, each value comprising a plurality of characters,
- wherein a respective plurality of values are generated for each field of the plurality of fields, wherein each value of said respective plurality of values is generated for the field by said performing character recognition using a corresponding approach of said plurality of approaches;
- for each field of said plurality of fields, selecting one value of said respective plurality of values for the field as representing a recognized best value for the field, wherein said selecting selects a first value generated by a first sequence of image processing techniques as the recognized best value for a first field and a second value generated by a second sequence of image processing techniques as the recognized best value for a second field, wherein said first field and said field are comprised in said plurality of fields;
- combining the respective recognized best value and said key for each field of said plurality of fields to form a result text representing the textual information contained in said image document;
- and
- providing said result text as a response to said request.
11 more elements
Dependencies are inferred from the claim text.