Methods and systems for determining authenticity of a document
A method for determining authenticity of a document is provided that includes receiving, by an electronic device, an image of a document, assigning a label to the image, and obtaining vectors for each image in a subset of images. Each image is of a document and is assigned the same label as the received image. Moreover, the method includes encoding the received image into a vector, calculating a distance between the vector of the received image and each obtained vector, comparing each of the calculated distances against a threshold distance, and calculating a number of the calculated distances that are less than or equal to the threshold distance. In response to determining the calculated number is at least equal to a required number, the document in the received image is determined to be authentic. Otherwise, the received image requires manual review.
1. A method for determining authenticity of a document comprising the steps of:
storing images of authentic documents in an image database as subsets of images defined based on the similarity of text content between the images;
receiving, by an electronic device, an image of a document;
assigning a label to the received image;
obtaining a low dimensionality vector for each image in one of the subsets, wherein each image in the one subset is encoded into the respective low dimensionality vector using a trained machine learning modelis and is assigned the same label as the received image;
encoding the received image into a low dimensionality vector;
calculating a distance between the low dimensionality vector of the received image and each obtained low dimensionality vector, each distance represents the similarity in appearance between the received image and a respective image in the one subset, wherein the smaller the distance the greater the similarity between the images;
comparing each of the calculated distances against a threshold distance;
calculating a number of the calculated distances that are less than or equal to the threshold distance;
in response to determining the calculated number is at least equal to a required number, determining the document in the received image is authentic; and
in response to determining the calculated number is less than the required number, determining the received image requires manual review.
2. The method according to claim 1 further comprising the step of determining the label for each image based on the text of the document included in the respective image.
3. The method according to claim 1 , said step of determining the received image requires manual review further comprising determining the document in the received image is not similar in content and appearance as the document in any image in the database.
4. The method according to claim 1 , said step of determining the document in the received image is authentic further comprising determining the document in the received image is similar in content and appearance as the document in a required number of images in the database.
5. The method according to claim 1 , further comprising training a machine learning model to encode an image of a document into a low dimensionality vector.
6. An electronic device for determining authenticity of a document comprising:
a processor; and
a memory configured to store data, said electronic device being associated with a network and said memory being in communication with said processor and having instructions stored thereon which, when read and executed by said processor, cause said electronic device to:
store images of authentic documents in an image database as subsets of images defined based on the similarity of text content between the images;
receive an image of a document;
assign a label to the received image;
obtain a low dimensionality vector for each image in one of the subsets, wherein each image in the one subset is encoded into the respective low dimensionality vector using a trained machine learning model and is assigned the same label as the received image;
encode the received image into a low dimensionality vector;
calculate a distance between the low dimensionality vector of the received image and each obtained low dimensionality vector, each distance represents the similarity in appearance between the received image and a respective image in the one subset, wherein the smaller the distance the greater the similarity between the images;
compare each of the calculated distances against a threshold distance;
calculate a number of the calculated distances that are less than or equal to the threshold distance;
in response to determining the calculated number is at least equal to a required number, determining the document in the received image is authentic; and
in response to determining the calculated number is less than the required number, determining the received image requires manual review.
7. The electronic device according to claim 6 , wherein the instructions when read and executed by said processor, cause said electronic device to determine the label for each image based on the text of the document included in the respective image.
8. The electronic device according to claim 6 , wherein the instructions when read and executed by said processor, cause said electronic device to determine the document in the received image is not similar in content and appearance as the document in any image in the database when manual review is required.
9. The electronic device according to claim 6 , wherein the instructions when read and executed by said processor, cause said electronic device to determine the document in the received image is similar in content and appearance as the document in a required number of images in the database when the document in the received image is determined to be authentic.
10. The electronic device according to claim 6 , wherein the instructions when read and executed by said processor, cause said electronic device to train a machine learning model to encode an image of a document into a low dimensionality vector.
11. A non-transitory computer-readable recording medium in an electronic device for determining authenticity of a document, the non-transitory computer-readable recording medium storing instructions which when executed by a hardware processor cause the non-transitory recording medium to perform steps comprising:
storing images of authentic documents in an image database as subsets of images defined based on the similarity of text content between the images;
receiving an image of a document;
assigning a label to the received image;
obtaining a low dimensionality vector for each image in one of the subsets, wherein each image in the one subset is encoded into the respective low dimensionality vector using a trained machine learning model and is assigned the same label as the received image;
encoding the received image into a low dimensionality vector;
calculating a distance between the low dimensionality vector of the received image and each obtained low dimensionality vector, each distance represents the similarity in appearance between the received image and a respective image in the one subset, wherein the smaller the distance the greater the similarity between the images;
comparing each of the calculated distances against a threshold distance;
calculating a number of the calculated distances that are less than or equal to the threshold distance;
in response to determining the calculated number is at least equal to a required number, determining the document in the received image is authentic; and
in response to determining the calculated number is less than the required number, determining the received image requires manual review.
12. The non-transitory computer-readable recording medium according to claim 11 , wherein the instructions when read and executed by said processor, cause said non-transitory computer-readable recording medium to perform a step of determining the label for each image based on the text of the document included in the respective image.
13. The non-transitory computer-readable recording medium according to claim 11 , wherein the instructions when read and executed by said processor, cause said non-transitory computer-readable recording medium to perform a step of determining the document in the received image is not similar in content and appearance as the document in any image in the database when the received image requires manual review.
14. The non-transitory computer-readable recording medium according to claim 11 , wherein the instructions when read and executed by said processor, cause said non-transitory computer-readable recording medium to perform a step of determining the document in the received image is similar in content and appearance as the document in a required number of images in the database when the received image is determined to be authentic.
15. The non-transitory computer-readable recording medium according to claim 11 , wherein the instructions when read and executed by said processor, cause said non-transitory computer-readable recording medium to train a machine learning model to encode an image of a document into a low dimensionality vector.