Access control method and device based on face recognition, equipment and storage medium

Abstract

The invention relates to an access control method and device based on face recognition, equipment and a storage medium, and the access control method comprises the steps: collecting an access control area image; using a pre-trained face detection model to detect a face of the acquired access control area image; if the human face is detected, performing human face feature extraction on a human face region in the access control region image by utilizing a pre-trained convolutional neural network model to obtain a human face feature vector corresponding to the access control region image, wherein the layer number of convolutional layers of the convolutional neural network model is smaller than or equal to a preset layer number; comparing the face feature vector corresponding to the access control area image with a pre-stored face feature vector in a pre-constructed face database; and if a pre-stored face feature vector matched with the face feature vector corresponding to the access control area image exists in the face database, releasing the access control. The memory occupation of the two models is only hundreds of kB, so that the two models can be stored and calculated locally.

Description

technical field [0001] The present invention relates to the technical field of face recognition, in particular to an access control method, device, equipment and storage medium based on face recognition. Background technique [0002] With the popularity of intelligent equipment, access control equipment is also showing an increasingly high-end trend. From the original key unlocking method to card verification, and then to the more common password verification and fingerprint verification, the way of access control is more and more separated from foreign objects, and more dependent on the biological characteristics of people themselves, making the identity verification method more Simple, fast, safe and reliable. [0003] Face recognition is another identity verification mechanism applied to the access control system, which is simpler than fingerprint verification or iris verification. In related technologies, the more common face recognition technology uses deep learning m...

AI Technical Summary

Problems solved by technology

In related technologies, the more common face recognition technology uses deep learning methods to train neural networks to complete the training of face recognition models, and often uses complex models and various features and parameters. The model trained based on this is usually There are problems such as difficult deployment and complex calculations, so the access control terminal often cannot complete the identification alone, and must cooperate with the server to perform complex calculations, which affects the identification efficiency

Images