Face recognition detection method and system based on mobile terminal edge calculation

Abstract

The invention discloses a face recognition detection method and system based on mobile terminal edge calculation. The method comprises the following steps: extracting a picture in a video shot by mobile terminal monitoring equipment; integrating the sizes of the pictures by adopting an NCNN framework to obtain a picture pyramid, and outputting two characteristic spectrums through a neural networkto perform foreground and background classification and boundary frame regression to identify a face frame; adopting an MTCNN network model to extract a first face information feature vector; adding alabel to the first face information feature vector, and storing the first face information feature vector as matching data; and enabling a rear-end monitoring device to collect a real-time picture, extracts a second face information feature vector of the picture, calculating an Euclidean distance between the first face information feature vector and the second face information feature vector, comparing the Euclidean distance with a set distance threshold, and identifying face information and a corresponding label in the real-time picture. According to the invention, effective human faces canbe screened, and the calculation pressure is reduced for the rear end, so that the recognition speed is relatively high on the premise of ensuring a high recognition success rate.

Description

technical field [0001] The invention relates to the technical field of computer vision and pattern recognition, in particular to a method and system for face recognition and detection based on mobile terminal edge computing. Background technique [0002] At present, face recognition is one of the hot research issues in the field of computer vision and pattern recognition, and how to make the front-end collected images to the back-end for faster and efficient processing has become a key and difficult issue in the field of face recognition. [0003] In the existing technology, a deep convolutional neural network (CNN) is used to learn to map the image to the Euclidean space. The spatial distance is directly related to the image similarity: the spatial distance between different images of the same person is very small, and the images of different people have a large distance in space, which can be used for face verification, recognition and clustering. As a classic deep learni...

AI Technical Summary

Problems solved by technology

[0004] However, in face detection, several layers of networks are used in face detection and face recognition. First, the face is identified, and then the feature value of the face is extracted. There are two problems with the traditional method: ( 1) Acquisition devices such as cameras are slow to process faces in video streams; (2) When performing face matching, the faces need to be identified first, and then the face features have to be extracted, and repeated face feature extraction Steps increase computation time

Images