Why is Amazon's palm recognition method more convenient?

Amazon recently announced that it will expand its palm scanning payment system to Whole Foods Market. When shoppers use the system for the first time, they need to insert a credit card to link it to the user's palm. After that, users only need to place their hand on the scanning device to pay.

Facilities such as stores, libraries, hospitals, offices, etc. may require the ability to identify users in the facility. Traditional systems have several significant drawbacks, including susceptibility to fraud, speed, accuracy, and operational limitations. Therefore, systems that use only "something you have" for identification are prone to misuse. Biometric identification systems address this issue by using characteristics of a specific individual that are difficult or impossible to replicate or transmit. However, traditional biometric identification systems can introduce operational issues and can also exhibit significant delays in heavily used environments.

To this end, Amazon applied for an invention patent called "Contactless Biometric Identification System" (application number: 201980046968.9) on June 20, 2019, and the applicant was Amazon Technologies Company.

Figure 1 System for identifying users at a facility

FIG1 is a system for identifying a user at a facility. The facility includes a door 104, which in turn includes a movable barrier to control the movement of a user 102, and one or more scanners 106. The scanner is capable of capturing an image of a hand, the image being illuminated using infrared light having two or more specific polarizations, having different illumination patterns, etc., and can generate raw image data 110. Depending on the polarization used, the image produced by the scanner has first modality features or second modality features, wherein the first modality features include features close to or on the surface of the user's hand, and the second modality features include those features below the epidermis.

The raw image data includes one or more of the following types of images: A raw first modality image 112 ( 1 ) and a raw second modality image 114 ( 1 ) obtained by a camera with a polarizer. Also included is another raw first modality image 112 ( 2 ) and another raw second modality image 114 ( 2 ).

The identification system 120 uses the raw image data as input. The initial processing module 122 prepares the raw image data for further processing to produce processed image data. The sub-image module 124 uses the processed image data as input and divides the corresponding image into sub-images. The sub-image module generates sub-image feature data 126 for one or more sub-images of at least a portion of the processed image data.

Comparison module 128 uses the sub-image feature data and previously stored user feature data 130 to determine the identity of the user based on features present in the image of the user's hand. The current distinctive features of the feature vector are compared to reference distinctive features in the user feature data, a confidence value indicating the similarity between the current distinctive features and the reference distinctive features can be determined, and then identification data 132 can be generated. As the user moves around the facility, facility management system 134 can use the identification data to associate an identity with the user.

In short, Amazon's identification system patent identifies users based on a comparison of current distinctive features with previously stored reference distinctive features associated with a user identifier, thereby improving recognition accuracy and making the system more secure.

Amazon is the largest online e-commerce company in the United States and one of the earliest companies to start e-commerce on the Internet. It has now become a comprehensive service provider. Building a customer-centric service-oriented enterprise has become Amazon's development direction, and in the future Amazon will further enhance customer experience.