System and method for optical character recognition

• US 9,990,564 B2
• Filed: 03/30/2016
• Issued: 06/05/2018
• Est. Priority Date: 03/29/2016
• Status: Active Grant

• Current Assignee: Wipro Limited
• Original Assignee: Wipro Limited
• Inventors: Singh, Madhusudan, Ramanna, Ramprasad Kanakatte, Mannopantar, Raghottam
• Primary Examiner(s): Grant, II, Jerome
• Application Number:
  US15/085,348
• Publication Number:
  US 20170286803A1
• Time in Patent Office: 797 Days
• Field of Search: 382161
• CPC Class Codes: G06F 18/2178 based on feedback of a supervisor
• CPC Class Codes: G06F 18/2433 Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
• CPC Class Codes: G06F 18/41 Interactive pattern learning with a human teacher
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/19133 Interactive pattern learning with a human teacher
• CPC Class Codes: G06V 30/19167 Active pattern learning

Abstract

This disclosure relates to system and method for optical character recognition. In one embodiment, the method comprises providing an image data to a plurality of customized machine learning algorithms or various customized neural networks, configured to recognize a set of pre-defined characters. The method comprises presenting one or more suggestions for the character to the user in response to negative character recognition, and training a customized machine learning algorithm corresponding to the character if one of the suggestions is identified by the user. If the suggestions are rejected by the user, the method comprises prompting the user to identify the character and determining presence of the character in the set of pre-defined characters. The method further comprises training a customized machine learning algorithm corresponding to the character if the character is present, or dynamically creating a customized machine learning algorithm corresponding to the character if the character is not present.

Claims

1. A method for optical character recognition, the method comprising:

detecting, by a character recognition device, a presence of a character in an image data;

providing, by the character recognition device, the image data to a plurality of customized machine learning algorithms for character recognition, wherein each of the plurality of customized machine learning algorithms is configured to recognize a pre-defined character;

in response to a negative character recognition, presenting, by the character recognition device, one or more suggestions for the character to the user, wherein the one or more suggestions comprises characters corresponding to one or more of the plurality of customized machine learning algorithms whose outputs meet a pre-defined threshold criteria for suggestion but does not meet a pre-defined threshold criteria for recognition;

in response to the one or more suggestions being rejected by the user,

prompting, by the character recognition device, the user to identify the character;

determining, by the character recognition device, a presence of the character in a set of pre-defined characters;

in response to a positive presence, training, by the character recognition device, a customized machine learning algorithm corresponding to the character; and

in response to a negative presence, adding, by the character recognition device, the character in the set of pre-defined characters, and dynamically creating, by the character recognition device, a customized machine learning algorithm corresponding to the character; and

in response to one of the one or more suggestions being identified by the user, training, by the character recognition device, a customized machine learning algorithm corresponding to the character.

2. The method of claim 1, further comprising:

receiving an image;

pre-processing the image to correct an orientation of the image or to filter noise from the image; and

processing the image to generate the image data.

3. The method of claim 1, wherein each of the plurality of customized machine learning algorithms is a multi-layer feed-forward neural network.

4. The method of claim 1, wherein character recognition comprises:

applying a pre-defined threshold criteria for recognition to output of each of the plurality of customized machine learning algorithms; and

selecting the character corresponding to one of the plurality of customized machine learning algorithms whose output meets the pre-defined threshold criteria for recognition.

5. The method of claim 1, further comprising gathering the recognized or identified characters for subsequent processing to generate a text.

6. The method of claim 1, wherein dynamically creating the customized machine learning algorithm corresponding to the character comprises learning the character by deep learning mechanism.

7. The method of claim 1, wherein training the customized machine learning algorithm corresponding to the character comprises learning the character by transfer learning mechanism.

8. A system for optical character recognition, the system comprising:

at least one processor; and

a non-transitory computer-readable medium storing instructions that, when executed by the at least one processor, cause the at least one processor to perform operations comprising;

detecting a presence of a character in an image data;

providing the image data to a plurality of customized machine learning algorithms for character recognition, wherein each of the plurality of customized machine learning algorithms is configured to recognize a pre-defined character;

in response to a negative character recognition, presenting one or more suggestions for the character to the user, wherein the one or more suggestions comprises characters corresponding to one or more of the plurality of customized machine learning algorithms whose outputs meet a pre-defined threshold criteria for suggestion but does not meet a pre-defined threshold criteria for recognition;

in response to the one or more suggestions being rejected by the user,

prompting the user to identify the character;

determining a presence of the character in a set of pre-defined characters;

in response to a positive presence, training a customized machine learning algorithm corresponding to the character; and

in response to a negative presence, adding the character in the set of pre-defined characters and dynamically creating a customized machine learning algorithm corresponding to the character; and

in response to one of the one or more suggestions being identified by the user, training a customized machine learning algorithm corresponding to the character.

9. The system of claim 8, wherein the operations further comprise:

receiving an image;

pre-processing the image to correct an orientation of the image or to filter noise from the image; and

processing the image to generate the image data.

10. The system of claim 8, wherein each of the plurality of customized machine learning algorithms is a multi-layer feed-forward neural network.

11. The system of claim 8, wherein character recognition comprises:

applying a pre-defined threshold criteria for recognition to output of each of the plurality of customized machine learning algorithms; and

selecting the character corresponding to one of the plurality of customized machine learning algorithms whose output meets the pre-defined threshold criteria for recognition.

12. The system of claim 8, wherein the operations further comprise gathering the recognized or identified characters for subsequent processing to generate a text.

13. The system of claim 8, wherein dynamically creating the customized machine learning algorithm corresponding to the character comprises learning the character by deep learning mechanism.

14. The system of claim 8, wherein training the customized machine learning algorithm corresponding to the character comprises learning the character by transfer learning mechanism.

15. A non-transitory computer-readable storage medium having stored thereon, a set of computer-executable instructions for causing a computer comprising one or more processors to perform steps comprising:

detecting a presence of a character in an image data;

providing the image data to a plurality of customized machine learning algorithms for character recognition, wherein each of the plurality of customized machine learning algorithms is configured to recognize a pre-defined character;

in response to a negative character recognition, presenting one or more suggestions for the character to the user, wherein the one or more suggestions comprises characters corresponding to one or more of the plurality of customized machine learning algorithms whose outputs meet a pre-defined threshold criteria for suggestion but does not meet a pre-defined threshold criteria for recognition;

in response to the one or more suggestions being rejected by the user,

prompting the user to identify the character;

determining a presence of the character in a set of pre-defined characters;

in response to a positive presence, training a customized machine learning algorithm corresponding to the character; and

in response to a negative presence, adding the character in the set of pre-defined characters and dynamically creating a customized machine learning algorithm corresponding to the character; and

in response to one of the one or more suggestions being identified by the user, training a customized machine learning algorithm corresponding to the character.

16. The non-transitory computer-readable medium of claim 15, further storing computer-executable instructions for:

receiving an image;

pre-processing the image to correct an orientation of the image or to filter noise from the image; and

processing the image to generate the image data.

17. The non-transitory computer-readable medium of claim 15, wherein character recognition comprises:

applying a pre-defined threshold criteria for recognition to output of each of the plurality of customized machine learning algorithms; and

selecting the character corresponding to one of the plurality of customized machine learning algorithms whose output meets the pre-defined threshold criteria for recognition.