Systems and methods for incremental character recognition to recognize characters in images
AssigneeFacebook, Inc.
InventorsBenjamin Ray, Ahmad Abdulmageed Mohammed Abdulkader, Sofus Attila Macskassy
Life of this patent
Estimating the term…
Abstract
Systems, methods, and non-transitory computer-readable media can acquire an image that depicts at least one character. A set of pixels, within the image, through which the at least one character is depicted can be identified. At least one linear portion, within the image, can be identified based on the set of pixels. For each sub-portion within the at least one linear portion, a respective first confidence score representing a respective first likelihood that a respective sub-portion depicts the at least one character can be determined.
Claims20
A computer-implemented method comprising:
- acquiring, by a computing system, an image that depicts at least one character;
- identifying, by the computing system, a set of pixels, within the image, through which the at least one character is depicted, wherein the identifying the set of pixels utilizes an image patch;
- identifying, by the computing system, at least one linear portion, within the image, based on the set of pixels, wherein the identifying the at least one linear portion utilizes a text heat-map;
- determining, by the computing system, for each sub-portion within the at least one linear portion, a respective first confidence score representing a respective first likelihood that a respective sub-portion depicts the at least one character;
- combining, by the computing system, the respective first confidence score and a respective second confidence score to produce a respective overall confidence score for the respective sub-portion;
- determining, by the computing system, that the respective overall confidence score at least meets a specified threshold value;
- and
- recognizing, by the computing system, the respective sub-portion as being associated with the at least one character.
A system comprising:
- at least one processor;
- and
- a memory storing instructions that, when executed by the at least one processor, cause the system to perform:
- acquiring an image that depicts at least one character;
- identifying a set of pixels, within the image, through which the at least one character is depicted, wherein the identifying the set of pixels utilizes an image patch;
- identifying at least one linear portion, within the image, based on the set of pixels, wherein the identifying the at least one linear portion utilizes a text heat-map;
- determining, for each sub-portion within the at least one linear portion, a respective first confidence score representing a respective first likelihood that a respective sub-portion depicts the at least one character;
- combining the respective first confidence score and a respective second confidence score to produce a respective overall confidence score for the respective sub-portion;
- determining that the respective overall confidence score at least meets a specified threshold value;
- and
- recognizing the respective sub-portion as being associated with the at least one character.
A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:
- acquiring an image that depicts at least one character;
- identifying a set of pixels, within the image, through which the at least one character is depicted, wherein the identifying the set of pixels utilizes an image patch;
- identifying at least one linear portion, within the image, based on the set of pixels, wherein the identifying the at least one linear portion utilizes a text heat-map;
- and
- determining, for each sub-portion within the at least one linear portion, a respective first confidence score representing a respective first likelihood that a respective sub-portion depicts the at least one character;
- combining the respective first confidence score and a respective second confidence score to produce a respective overall confidence score for the respective sub-portion;
- determining that the respective overall confidence score at least meets a specified threshold value;
- and
- recognizing the respective sub-portion as being associated with the at least one character.
Dependencies are inferred from the claim text.