Document optical character recognition
Vehicles and other items often have corresponding documentation, such as registration cards, that includes a significant amount of informative textual information that can be used in identifying the item. Traditional OCR may be unsuccessful when dealing with non-cooperative images. Accordingly, features such as dewarping, text alignment, and line identification and removal may aid in OCR of non-cooperative images. Dewarping involves determining curvature of a document depicted in an image and processing the image to dewarp the image of the document to make it more accurately conform to the ideal of a cooperative image. Text alignment involves determining an actual alignment of depicted text, even when the depicted text is not aligned with depicted visual cues. Line identification and removal involves identifying portions of the image that depict lines and removing those lines prior to OCR processing of the image.
1 . A method comprising:
receiving, from a client device via a network, an image of a document comprising text;
identifying one or more curved lines in the image;
applying a dewarping function to the image based on the one or more curved lines in the image, the dewarping function comprising a transformation that straightens the one or more curved lines in the image;
performing image analysis on the dewarped image;
based on the image analysis, identifying a known document, from a database of a plurality of known documents, that best matches the dewarped image of the document;
using the known document as a template, identifying a location within the dewarped image that contains particular text by mapping regions of the dewarped image to document-specific fields within the known document;
performing optical character recognition within the identified location to extract the particular text; and
causing presentation of the extracted particular text.
2 . The method of claim 1 , further comprising:
identifying a color of a line segment depicted in the image, wherein the optical character recognition is performed on portions of the image depicting the particular text while ignoring elements within the portions having the color of the line segment.
3 . The method of claim 1 , wherein performing the image analysis comprises identifying spacing of lines of text in the image.
4 . The method of claim 3 , wherein:
identifying the spacing of lines of text comprises virtually superimposing grids of lines over the text with different positions and spacings; and
a position and spacing of a grid that overlaps the least with identified locations of text is used as a basis for text alignment.
5 . The method of claim 3 , wherein performing the image analysis further comprises identifying page edges in the image.
6 . The method of claim 5 , further comprising:
identifying portions of the dewarped image that depict text based on the identified page edges and the identified spacing of lines.
7 . The method of claim 1 , wherein identifying the known document comprises matching based on one or more of an aspect ratio, a number of lines, a line spacing to page height ratio, or a watermark.
8 . The method of claim 1 , further comprising:
accessing, from a second database, additional information regarding an item based on the extracted particular text; and
generating a listing for the item by including at least some of the additional information accessed from the second database.
9 . A system comprising:
one or more processors; and
a memory storing instructions which, when executed by the one or more processors, cause the system to perform operations comprising:
receiving, from a client device via a network, an image of a document comprising text;
identifying one or more curved lines in the image;
applying a dewarping function to the image based on the one or more curved lines in the image, the dewarping function comprising a transformation that straightens the one or more curved lines in the image;
performing image analysis on the dewarped image;
based on the image analysis, identifying a known document, from a database of a plurality of known documents, that best matches the dewarped image of the document;
using the known document as a template, identifying a location within the dewarped image that contains particular text by mapping regions of the dewarped image to document-specific fields within the known document;
performing optical character recognition within the identified location to extract the particular text; and
causing presentation of the extracted particular text.
10 . The system of claim 9 , wherein the operations further comprise:
identifying a color of a line segment depicted in the image, wherein the optical character recognition is performed on portions of the image depicting the particular text while ignoring elements within the portions having the color of the line segment.
11 . The system of claim 9 , wherein performing the image analysis comprises identifying spacing of lines of text in the image.
12 . The system of claim 11 , wherein:
the identifying the spacing of lines of text comprises virtually superimposing grids of lines over the text with different positions and spacings; and
a position and spacing of a grid that overlaps the least with identified locations of text is used as a basis for text alignment.
13 . The system of claim 11 , wherein performing the image analysis further comprises identifying page edges in the image.
14 . The system of claim 13 , wherein the operations further comprise
identifying portions of the dewarped image that depict text based on the identified page edges and the identified spacing of lines.
15 . The system of claim 9 , wherein identifying the known document comprises matching based on one or more of an aspect ratio, a number of lines, a line spacing to page height ratio, or a watermark.
16 . The system of claim 9 , wherein the operations further comprise:
accessing, from a second database, additional information regarding an item based on the extracted particular text; and
generating a listing for the item by including at least some of the additional information accessed from the second database.
17 . A non-transitory computer-readable medium storing executable instructions which, when executed by one or more processors of a server, cause the server to perform operations comprising:
receiving, from a client device via a network, an image of a document comprising text;
identifying one or more curved lines in the image;
applying a dewarping function to the image based on the one or more curved lines in the image, the dewarping function comprising a transformation that straightens the one or more curved lines in the image;
performing image analysis on the dewarped image;
based on the image analysis, identifying a known document, from a database of a plurality of known documents, that best matches the dewarped image of the document;
using the known document as a template, identifying a location within the dewarped image that contains particular text by mapping regions of the dewarped image to document-specific fields within the known document;
performing optical character recognition within the identified location; and
causing presentation of the extracted particular text.
18 . The computer-readable medium of claim 17 , wherein the operations further comprise:
identifying a color of a line segment depicted in the image, wherein the optical character recognition is performed on portions of the image depicting the particular text while ignoring elements within the portions having the color of the line segment.
19 . The computer-readable medium of claim 17 , wherein identifying the known document comprises matching based on one or more of an aspect ratio, a number of lines, a line spacing to page height ratio, or a watermark.
20 . The computer-readable medium of claim 17 , wherein the operations further comprise:
accessing, from a second database, additional information regarding an item based on the extracted particular text; and
generating a listing for the item by including at least some of the additional information accessed from the second database.