Customer identification verification
A transaction terminal captures at least one image of a customer during a checkout and the customer is asked to place a photo identification card in view of the camera where a second image is captured of the card. The second image is converted to text and various components of the card are validated. Any hologram on the card is verified from the second image. The photo on the card is compared to facial features in the customer image and a determination is made as to whether the customer's identification can or cannot be verified. When verified, a message is sent to an attendant's device with the determination that the customer and card were verified along with the customer image and card image. The attendant is asked to confirm the verification and when confirmed the terminal resumes processing the checkout on behalf of the customer.
1 . A method, comprising:
receiving, by a transaction terminal equipped with at least one processor and a non-transitory computer-readable storage medium, a first image of a face of a customer who is present at the transaction terminal during a checkout and receiving a second image of a photo identification card provided by the customer at the transaction terminal;
wherein receiving includes:
receiving the first image and the second image together in a single image that is segmented into the first image and the second image, wherein the single image is captured by a camera of the transaction terminal;
comparing, using a machine-learning model executed by a processor of a server, first features of the face against second features from a photo of the customer's face provided in a portion of the second image;
wherein comparing includes:
passing the first image and the second image to the machine-learning model as input and receiving a percentage facial match as output from the machine-learning model;
validating, via optical character recognition software executed by the at least one processor, text information for the photo identification card derived from the second image, wherein validating includes checking security features specific to a type of the photo identification card; and
determining, based on the comparing and the validating, a confidence percentage as to whether an identification of the customer can be verified based on the comparing and the validating, wherein the determining includes employing an algorithm tailored to enhance security in real time.
2 . The method of claim 1 , further comprising sending the first image, the second image, and the confidence percentage to a management terminal that is monitoring the checkout at the transaction terminal.
3 . The method of claim 2 , further comprising:
receiving a decision from the management terminal indicating whether the identification of the customer: was accepted based on the confidence percentage, was rejected based on the confidence percentage, or was overridden by an operator of the management terminal; and
logging the first image, the second image, the confidence percentage, and the decision in a compliance log.
4 . The method of claim 1 , wherein receiving further includes obtaining the first image and the second image when the transaction terminal interrupts the checkout for a customer age-check because of an age-restricted item code scanned during the checkout.
5 . The method of claim 1 , further comprising cropping the face from the first image into a facial image of a first size, scaling the portion of the second image to the first size as a scaled facial image, extracting the first features from the facial image, and extracting the second features from the scaled facial image as part of the comparing.
6 . The method of claim 5 , wherein cropping further includes pre-processing the scaled facial image by adjusting brightness and color attributes in the scaled facial image before the comparing.
7 . The method of claim 6 , wherein pre-processing further includes passing the facial image and the scaled facial image to the machine-learning model and receiving the percentage facial match as output from the machine-learning model.
8 . The method of claim 7 , wherein validating further includes verifying parameters, text derived information from a second portion of the second image, and image security features in the second portion of the second image to generate a real card percentage indicating a likelihood the photo identification card is a legitimate card.
9 . The method of claim 8 , wherein determining further includes processing rules to generate the confidence percentage from the percentage facial match and the real card percentage.