Method and system for adaptively improving face image acquisition and storage medium

Abstract

The invention discloses a method and a system for adaptively improving face image acquisition and a storage medium. The method comprises the following steps: acquiring attitude information, coordinateinformation and size information of a face in a picture; according to the coordinate information and the size information of the human face, calculating the proportion information of the human face area and the preview image area, and determining the recognition distance between the human face and the camera device; dynamically adjusting identification parameters according to the identification distance; and based on the adjusted recognition parameters, collecting a face image through camera equipment. According to the method for dynamically adjusting the recognition parameters of the cameradevice, it can be guaranteed that the collected face image is in the center of the preview image, the recognition distance between the face and the camera device meets the recognition requirement, theliving body recognition and comparison recognition accuracy is improved, the application range is wide, and the method can be widely applied to the technical field of face recognition.

Description

technical field [0001] The invention relates to the technical field of face recognition, in particular to a method, system and storage medium for adaptively improving face image acquisition. Background technique [0002] Face liveness detection is a method to determine the real physiological characteristics of an object in some authentication scenarios. In face recognition applications, liveness detection can be realized in the monocular camera through combined actions such as blinking, opening the mouth, shaking the head, and nodding. The binocular camera is realized in a silent way, that is, it does not require the customer to cooperate with the action, and uses technologies such as face key point positioning and face tracking to verify whether the user is a real living person. This method can effectively resist common attack methods such as photos, face changes, masks, occlusions, and screen recaptures, thereby helping users identify fraudulent behavior and protecting use...

AI Technical Summary

Problems solved by technology

[0004]1) Bank self-service machines are aimed at ordinary customers, and the heights of customers vary. Currently, binocular cameras generally have a vertical viewing angle of 55°, but they are still not effective When identifying people whose height is around 1.4 meters and 1.9 meters, the face imaging of this part of the crowd is often at the top or bottom of the preview screen, or even only part of the face is on the preview screen, and the pitch angle of the face is too high. Large, affecting the recognition effect

[0005]2), customers often have randomness in the process of identifying and taking photos, and the collected face images are on the far left or right of the preview screen, resulting in some faces Not collected, resulting in the loss of some facial features; in addition, the left and right deflection of the face is too large, resulting in a left and right deflection angle of the collected face, which affects the face recognition and face comparison

[0006]3) In the face recognition process, the distance between the face and the camera is generally required to be within the range of 30cm---80cm, while in the bank non-cash counter equipment In this case, the vertical distance between the installation position of the camera and the edge of the device is about 50cm; while the customer generally keeps a distance of 30-80cm from the edge of the device during the transaction process, resulting in the longest distance between the face and the camera reaching about 130cm , unable to meet the distance requirements of existing face recognition equipment

When the distance is too far, the image is too dark; when the distance is too close, the image is overblown; and such an imaging effect leads to a high failure rate of face recognition

[0007]4), not only in bank self-service equipment, but also in the fields of public security, attendance and access control, etc., the above objective environmental problems, and binocular cameras are generally With a fixed focal length, face image acquisition is prone to blurring, resulting in low accuracy of face recognition and comparison recognition, and poor customer experience

Images