基于深度学习的实时人脸检测  

Real-time face detection based on deep learning

作　　者：郭昕刚[1] 屈诺希 杨洛 GUO Xingang;QU Nuoxi;YANG Luo(School of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China)

机构地区：[1]长春工业大学计算机科学与工程学院,吉林长春130012

出　　处：《长春工业大学学报》2021年第2期181-185,共5页Journal of Changchun University of Technology

基　　金：吉林省发改委基金资助项目(2019C040-3)。

摘　　要：利用Retinaface神经网络和WIDER FACE人脸数据集训练人脸检测模型,采用模型压缩剪枝方法对原有模型进行压缩,使其减少参数量,实现实时运行,然后将压缩后的模型部署到FPGA上。进行了相关实验。Retinaface neural network and WIDER FACE face data set are sued to train a face detection model.We apply he model compression pruning method to compress the original model for reducing the number of parameters to realize real-time operation.The compressed model is installed into FPGA.Experiments are carried out.

关 键 词：FPGA Retinaface 人脸检测 模型压缩 WIDER FACE数据集 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]