基于多任务卷积神经网络的实时人眼检测方法  

Real-Time Eye Detection Based on Multi-Task Convolutional Neural Networks

作　　者：张成 陈杰春[2] 吴猛[1] 陈旭 ZHANG Cheng;CHEN Jiechun;Wu Meng;Chen Xu(School of Information and Control Engineering,Jilin Institute of Chemical Technology,Jinlin Jilin 132022,China;School of Automation Engineering,Northeast Dianli University,Jilin Jilin 132012,China)

机构地区：[1]吉林化工学院信息与控制工程学院,吉林吉林132022 [2]东北电力大学自动化工程学院,吉林吉林132012

出　　处：《信息与电脑》2022年第17期83-85,共3页Information & Computer

摘　　要：针对戴眼镜、人脸姿态变化以及眯眼睛等复杂场景,提出了一种基于多任务卷积神经网络(MultiTask CascadedConvolutionalNetworks,MTCNN)的人眼检测算法。针对性地调整与优化网络,删除landmark部分以简化网络结构,进而调整网络的输入尺寸,使模型更适用于人眼检测。实验结果表明,基于MTCNN的人眼检测算法在数据集上准确率达92.1%,图形处理器(GraphicsProcessingUnit,GPU)检测速度达112frames/s,可以有效兼顾实时性与准确性。For complex scenes such as wearing glasses,face pose changes and squinting,this paper proposes a human eye detection algorithm based on MTCNN(Multi Task Cascaded Convolutional Networks).In this paper,the network is adjusted and optimized,the landmark part is deleted to simplify the network structure,the input size of the network is adjusted to make the model more suitable for human eye detection,and the structure of the model is adjusted to improve the performance of human eye detection.The experimental results show that the accuracy of the algorithm proposed in this paper is 92.1%on data set,and the Graphics Processing Unit(GPU)detection speed is 112 frames/s.It can effectively give consideration to real-time and accuracy.

关 键 词：人眼检测 多任务卷积神经网络(MTCNN) 图形处理器(GPU) 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.41[自动化与计算机技术—控制科学与工程]