offline multi-face recognition method based on Android equipment and computer equipment

Abstract

The invention discloses an offline multi-face recognition method based on Android equipment and computer equipment. The offline multi-face recognition method specifically comprises the steps that an interface module receives face imaging data collected by a camera in real time, and face imaging data are obtained; The service module dispatches the tool module to decode and convert the format into format data; The face detection module converts the format data into a face area coordinate set; The living body detection module filters non-living body data in the human face area coordinate set to obtain a living body human face data set, and the feature extraction module performs feature extraction on the format data and the living body human face data set to obtain a human face feature set; The face recognition module compares the face feature set with face feature data in the service module to obtain a similarity result and returns the similarity result to the service module; And the service module outputs the current face and the similarity result to the user interface through the interface module. According to the invention, there is no need to introduce too many additional sensors,and rapid face recognition can be realized under the condition of low equipment performance; The method can be applied to the scenes of face unlocking, face access control, face card punching and thelike.

Description

technical field [0001] The present invention specifically relates to an offline multi-face recognition method based on an Android device and a computer device. Background technique [0002] At present, the mainstream face recognition on the Android system is divided into two types: offline and online. Due to the limitations of device performance, camera and other hardware, the offline solution cannot achieve a fast enough recognition speed under the premise of meeting commercial accuracy, especially in simultaneous recognition. It is especially serious when there are multiple faces and a single face matches multiple registered faces. The offline face recognition system in the current field has high requirements on the camera. This is because the traditional face recognition is based on the Histogram of Oriented Gradients (HOG) feature, which requires high image resolution and poor generalization ability. When the image resolution is small, there will be a problem that the f...

AI Technical Summary

Problems solved by technology

[0002] At present, the mainstream face recognition on the Android system is divided into two types: offline and online. Due to the limitations of device performance, camera and other hardware, the offline solution cannot achieve a fast enough recognition speed under the premise of meeting commercial accuracy, especially in simultaneous recognition. It is especially serious when there are multiple faces and a single face matches multiple registered faces

The offline face recognition system in the current field has high requirements on the camera. This is because the traditional face recognition is based on the Histogram of Oriented Gradients (HOG) feature, which requires high image resolution and poor generalization ability. When the image resolution is small, there will be a problem of not being able to recognize the face, which will undoubtedly increase the hardware cost of the device

Images