Back to results
US 12,518,347 B2Granted
Systems and methods for identifying and correcting blurred components within images
AssigneeCapital One Services, LLC
InventorsJoseph James Ford, III, Thomas Oliver Cantrell
Life of this patent
Estimating the term…
Abstract
Methods and systems are described herein for identifying the location and nature of any blur within one or more images received as a user communication and generating an appropriate correction. The system utilizes a first machine learning model, which is trained to identify blurred components of inputted images and determine whether the blurred components are located in portions of the inputted images comprising textual information. The system may apply a corrective action selected by the first machine learning model, which may comprise stitching blurred images together to a sharp product image and/or some other method appropriate for rectifying images received.
Claims18
A system for identifying blurring of electronically submitted images during optical document recognition received as part of a client communication, the system comprising:
- one or more processors;
- and
- one or more non-transitory, computer-readable media comprising instructions that when executed by the one or more processors cause operations comprising:
- receiving, from a user, a user communication comprising a first image depicting subject matter, wherein one or more first portions of the first image comprises textual information and one or more second portions of the first image comprise non-textual information, and wherein the textual information comprises one or more alphanumeric characters, and wherein the non-textual information comprises one or more graphical portions;
- generating a feature input based on the user communication;
- inputting the feature input into a first machine learning model that performs the optical document recognition, wherein the first machine learning model is trained to (i) identify blurred components of inputted images, (ii) determine whether the blurred components are located in portions of the inputted images comprising textual information, and (iii) classify a cause of the blurred components to identify an error type;
- receiving an output from the first machine learning model;
- determining, based on the output, that (i) the first image has a blurred component, (ii) the blurred component is located within the one or more first portions of the first image comprising the textual information such that the blurred component causes the one or more alphanumeric characters to be obfuscated during the optical document recognition, and (iii) an error type indicating a cause of the blurred component;
- selecting, based on the error type, a corrective action from a plurality of corrective actions;
- and
- performing the corrective action to the first image to generate a second image depicting the subject matter comprising the one or more first portions comprising the textual information and the one or more second portions comprising the non-textual information, wherein the blurred component is absent from the textual information of the second image, wherein performing the corrective action comprises:
- retrieving a second machine learning model trained to stitch images together to correct blurring;
- and
- processing the first image using the second machine learning model to generate the second image.
11 more elements
A method for comprising:
- receiving, from a user, a user communication comprising a first image of subject matter, wherein the first image comprises textual information and non-textual information;
- generating a feature input based on the user communication;
- inputting the feature input into a first machine learning model trained to (i) identify blurred components of inputted images (ii) determine whether the blurred components are located in portions of the inputted images comprising textual information, and (iii) classify a cause of the blurred components to identify an error type;
- receiving an output from the first machine learning model;
- determining, based on the output, that (i) the first image has a first blurred component, (ii) the first blurred component is located at a portion of the first image comprising the textual information, and (iii) an error type indicating a cause of the first blurred component;
- selecting, based on the error type, a corrective action from a plurality of corrective actions;
- performing the corrective action to the first image;
- and
- generating, based on the corrective action being performed to the first image, a second image of the subject matter comprising the textual information and the non-textual information without the first blurred component.
6 more elements
One or more non-transitory, computer-readable media, comprising instructions that, when executed by one or more processors, cause operations comprising:
- receiving, from a user, a user communication comprising a first image comprising textual information and non-textual information;
- generating a feature input based on the user communication;
- inputting the feature input into a first machine learning model, trained to (i) identify blurred components of inputted images, (ii) determine whether the blurred components are located in portions of the inputted images comprising textual information, and (iii) classify a cause of the blurred components;
- receiving an output from the first machine learning model;
- determining, based on the output that (i) the first image has a first blurred component, (ii) the first blurred component is located at a portion of the first image comprising the textual information, and (iii) an error type indicating a cause of the first blurred component;
- selecting, based on the error type, a corrective action from a plurality of corrective actions;
- and
- performing the corrective action to the first image to generate a second image comprising the textual information and the non-textual information without the first blurred component.
5 more elements
Dependencies are inferred from the claim text.