Bank client identity authentication method and device

Abstract

The invention discloses a bank client identity verification method and device. The method comprises the steps: obtaining onsite pictures and identity document information used by different transactions from a transaction information retention record; feature value comparison is performed on the scene photos used for different transactions in pairs, and the similarity of different scene photo combinations is determined; determining whether each transaction is abnormal or not according to the similarity of different scene photo combinations and a preset similarity threshold value; inputting theidentity document information used by the abnormal transaction and the characteristic value of the on-site photo into a blacklist; and during the transaction, matching the onsite photo and the identity document information used in the transaction with the blacklist, and verifying the identity of the client. According to the invention, the blacklist can be determined according to the transaction information retention record and is used for carrying out identity verification on the bank client, so that the accuracy of identity verification of the bank client is improved, the situation that criminals still pass recognition by illegally using identities of others for many times is avoided, and financial risks are reduced.

Description

technical field [0001] The invention relates to the technical field of face recognition, in particular to a method and device for bank customer identity verification. Background technique [0002] This section is intended to provide a background or context to embodiments of the invention that are recited in the claims. The descriptions herein are not admitted to be prior art by inclusion in this section. [0003] At present, many transaction scenarios within banks use face recognition for customer identity verification, such as smart counters and mobile banking. When performing face recognition, it is necessary to compare the customer's on-site photo with a reference photo (such as an ID card photo). Among them, the on-site photo is the photo taken on the spot during the actual transaction, and generally is the best face photo intercepted in the liveness detection video stream; the reference photo is the ID photo used to compare with the on-site photo, usually the ID card ...

AI Technical Summary

Problems solved by technology

[0004] Traditional technologies usually use reference photos when authenticating bank customers. The reference photos are generally ID card photos, and the reference photos are generally old. Therefore, the face recognition results for the reference photos under the traditional technology are not perfect. It is 100% accurate, which leads to the fact that non-personal on-site photos may also pass face recognition during face recognition; moreover, in actual transactions, there are indeed cases where criminals have repeatedly used other people's identities to pass recognition, so traditional technology is Authentication of bank customers is not 100% accurate, so there is some financial risk involved

Images