Document image quality assessment
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for assessing document image quality are disclosed. In one aspect, a method includes the actions of receiving one or more images of a document. The actions further include detecting an object that has a shape that is similar to an expected shape of the document. The actions further include determining a focus score that reflects a sharpness of the first image. The actions further include determining a border fit score that reflects a likelihood that the object corresponds to the document. The actions further include determining an expected shape score that reflects a correlation between the shape of the object and an expected shape of the document. The actions further include determining a quality score that reflects a quality of the first image. The actions further include determining whether to perform document authentication on the first image.
1. A computer-implemented method comprising:
receiving, by a computing device, one or more images of a document;
detecting, by the computing device and in a first image of the one or more images, an object that has a shape that is similar to an expected shape of the document;
determining, by the computing device, a focus score that reflects a sharpness of the first image;
determining, by the computing device, a level of contrast between a border of the object and a portion of the image surrounding the object;
based on the level of contrast between the border of the object and the portion of the image surrounding the object, determining, by the computing device, a border fit score that reflects a likelihood that the border of the object is a border of the document;
determining, by the computing device, an angle defined by adjacent edges of the border of the object;
comparing, by the computing device, the angle defined by the adjacent edges of the border of the object to an expected angle of adjacent edges of the document;
based on comparing the angle defined by the adjacent edges of the border of the object to an expected angle of adjacent edges of the document, determining, by the computing device, an expected shape score that reflects a correlation between the shape of the object and an expected shape of the document;
based on the focus score, the border fit score, and the expected shape score, determining, by the computing device, a quality score that reflects a quality of the first image;
selecting, by the computing device, from among multiple quality score thresholds for each of multiple types of computing devices, a quality score threshold based on a type of the computing device;
determining, by the computing device, whether the quality score satisfies the quality score threshold; and
based on determining whether the quality score satisfies the quality score threshold, determining, by the computing device, whether to perform document authentication on the first image.
2. The method of claim 1 , wherein:
determining whether the quality score satisfies the quality score threshold comprises determining that the quality score satisfies the quality score threshold, and
determining whether to perform document authentication on the first image comprises determining to perform document authentication on the first image based on determining that the quality score satisfies the quality score threshold.
3. The method of claim 1 , wherein:
determining whether the quality score satisfies the quality score threshold comprises determining that the quality score does not satisfy the quality score threshold, and
determining whether to perform document authentication on the first image comprises determining not to perform document authentication on the first image based on determining that the quality score does not satisfy the quality score threshold.
4. The method of claim 3 , comprising:
in response to determining not to perform document authentication on the first image, determining a quality score for a second image of the one or more images.
5. The method of claim 1 , comprising:
determining whether the focus score satisfies a focus score threshold,
wherein determining whether to perform document authentication on the first image is further based on determining whether the focus score satisfies the focus score threshold.
6. The method of claim 1 , comprising:
determining whether the border fit score satisfies a border fit score threshold,
wherein determining whether to perform document authentication on the first image is further based on determining whether the border fit score satisfies the border fit score threshold.
7. The method of claim 1 , comprising:
determining whether the expected shape score satisfies an expected shape score threshold,
wherein determining whether to perform document authentication on the first image is further based on determining whether the expected shape score satisfies the expected shape score threshold.
8. The method of claim 1 , wherein the focus score reflects a sharpness of only the object.
9. The method of claim 1 , wherein the document is an identification document.
10. The method of claim 1 , wherein the one or more images of the document are frames of a video.
11. The method of claim 1 , comprising:
determining, for each of the one or more images of the document, a quality score; and
determining that the quality score of the first image is greater than the quality scores of each of the one or more images other than the first image,
wherein determining whether the quality score of the first image satisfies a quality score threshold comprises determining that the quality score of the first image satisfies the quality score threshold, and
wherein determining whether to perform document authentication on the first image comprises determining to perform document authentication on the first image based on determining that the quality score of the first image satisfies the quality score threshold and based on determining that the quality score of the first image is greater than the quality scores of each of the one or more images other than the first image.
12. The method of claim 1 , wherein:
determining whether the quality score satisfies a quality score threshold comprises determining that the quality score does not satisfy the quality score threshold, and
based on determining that the quality score does not satisfy the quality score threshold the method further comprises providing, to the user, feedback for improving quality of an image of the document.
13. A system comprising:
one or more computers; and
one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
receiving, by a computing device, one or more images of a document;
detecting, by the computing device and in a first image of the one or more images, an object that has a shape that is similar to an expected shape of the document;
determining, by the computing device, a focus score that reflects a sharpness of the first image;
determining, by the computing device, a level of contrast between a border of the object and a portion of the image surrounding the object;
based on the level of contrast between the border of the object and the portion of the image surrounding the object, determining, by the computing device, a border fit score that reflects a likelihood that the border of the object is a border of the document;
determining, by the computing device, an angle defined by adjacent edges of the border of the object;
comparing, by the computing device, the angle defined by the adjacent edges of the border of the object to an expected angle of adjacent edges of the document;
based on comparing the angle defined by the adjacent edges of the border of the object to an expected angle of adjacent edges of the document, determining, by the computing device, an expected shape score that reflects a correlation between the shape of the object and an expected shape of the document;
based on the focus score, the border fit score, and the expected shape score, determining, by the computing device, a quality score that reflects a quality of the first image;
selecting, by the computing device, from among multiple quality score thresholds for each of multiple types of computing devices, a quality score threshold based on a type of the computing device;
determining, by the computing device, whether the quality score satisfies the quality score threshold; and
based on determining whether the quality score satisfies the quality score threshold, determining, by the computing device, whether to perform document authentication on the first image.
14. The system of claim 13 , wherein:
determining whether the quality score satisfies the quality score threshold comprises determining that the quality score satisfies the quality score threshold, and
determining whether to perform document authentication on the first image comprises determining to perform document authentication on the first image based on determining that the quality score satisfies the quality score threshold.
15. The system of claim 13 , wherein:
determining whether the quality score satisfies the quality score threshold comprises determining that the quality score does not satisfy the quality score threshold, and
determining whether to perform document authentication on the first image comprises determining not to perform document authentication on the first image based on determining that the quality score does not satisfy the quality score threshold.
16. The system of claim 13 , wherein the operations further comprise:
determining whether the focus score satisfies a focus score threshold,
wherein determining whether to perform document authentication on the first image is further based on determining whether the focus score satisfies the focus score threshold.
17. The system of claim 13 , wherein the operations further comprise:
determining whether the border fit score satisfies a border fit score threshold,
wherein determining whether to perform document authentication on the first image is further based on determining whether the border fit score satisfies the border fit score threshold.
18. The system of claim 13 , wherein the operations further comprise:
determining whether the expected shape score satisfies an expected shape score threshold,
wherein determining whether to perform document authentication on the first image is further based on determining whether the expected shape score satisfies the expected shape score threshold.
19. The system of claim 13 , wherein the operations further comprise:
determining, for each of the one or more images of the document, a quality score; and
determining that the quality score of the first image is greater than the quality scores of each of the one or more images other than the first image,
wherein determining whether the quality score of the first image satisfies a quality score threshold comprises determining that the quality score of the first image satisfies the quality score threshold, and
wherein determining whether to perform document authentication on the first image comprises determining to perform document authentication on the first image based on determining that the quality score of the first image satisfies the quality score threshold and based on determining that the quality score of the first image is greater than the quality scores of each of the one or more images other than the first image.
20. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:
receiving, by a computing device, one or more images of a document;
detecting, by the computing device and in a first image of the one or more images, an object that has a shape that is similar to an expected shape of the document;
determining, by the computing device, a focus score that reflects a sharpness of the first image;
determining, by the computing device, a level of contrast between a border of the object and a portion of the image surrounding the object;
based on the level of contrast between the border of the object and the portion of the image surrounding the object, determining, by the computing device, a border fit score that reflects a likelihood that the border of the object is a border of the document;
determining, by the computing device, an angle defined by adjacent edges of the border of the object;
comparing, by the computing device, the angle defined by the adjacent edges of the border of the object to an expected angle of adjacent edges of the document;
based on comparing the angle defined by the adjacent edges of the border of the object to an expected angle of adjacent edges of the document, determining, by the computing device, an expected shape score that reflects a correlation between the shape of the object and an expected shape of the document;
based on the focus score, the border fit score, and the expected shape score, determining, by the computing device, a quality score that reflects a quality of the first image;
selecting, by the computing device, from among multiple quality score thresholds for each of multiple types of computing devices, a quality score threshold based on a type of the computing device;
determining, by the computing device, whether the quality score satisfies the quality score threshold; and
based on determining whether the quality score satisfies the quality score threshold, determining, by the computing device, whether to perform document authentication on the first image.