Human face acquisition quality evaluation strategy in community monitoring scene

Abstract

The invention relates to the technical field of face detection and deep learning, and particularly discloses a human face acquisition quality evaluation strategy in a community monitoring scene, whichcomprises the following steps: collecting an unconstrained pedestrian image set through a plurality of cameras in a community; training a target detection network Fast RCNN network into a human facedetection model to detect the position of a human face in the pedestrian image by utilizing a human face database wider face, and cutting the human face; calculating the image quality of the cropped face image from the two aspects of illuminance and definition through a traditional image processing method, and screening out high-quality and low-quality images in a manual screening mode to make a face image data set, wherein the obtained data set is used for training face quality evaluation network model FQANet parameters and the like. According to the invention, the practicability and applicability of an image quality evaluation strategy are improved by collecting an unconstrained data set, the accuracy of recognizing high-quality images by a network model designed in a community environment is about 95%, and the operation efficiency of the model is greatly improved.

Description

technical field [0001] The invention relates to the technical fields of face detection and deep learning, in particular to a face collection quality evaluation strategy in a community monitoring scene. Background technique [0002] In recent years, with the rapid development of machine learning and deep learning, various fields have been driven to realize technological changes. The application of pattern recognition technology in commercial, military, agricultural and other major fields is a research hotspot, and face recognition, as an important topic in the field of pattern recognition, is also a very concerned research direction, with strong application value and research significance. At present, face recognition technology has made great progress, but in practical applications, due to the influence of unfavorable factors such as illumination, posture, and low image pixels, the accuracy of face recognition is greatly reduced. Although many face recognition methods enha...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem of providing high-quality image sources for face recognition in the community monitoring scene, the purpose of the present invention is to provide a face collection quality evaluation strategy in the community monitoring scene

Images