一种用于多源人脸图像检索的损失函数改进策略

Improved loss function strategy for multi source face image retrieval

目前,公安机关刑侦部门大量的离线视频存储在监控网络的服务器中,为了在这些海量视频帧中提取嫌疑人目标人脸,设计了人脸检索系统。通过改变CNN网络的RELU结构,训练新的四分支孪生网络来获取深度特征,构造了一个新的四分支孪生网络。结合网上发布的逃犯人脸图片发起通缉,借助深度特征对比展开基于内容的图像检索(CBIR)。新的四分支孪生网络比熟悉的网络,如Alexnet、Googlenet、VGGNet和ResNet等收敛得更快,系统鲁棒性好。网络的共享权重设计使得检索具有较高的模型训练精度和检索精度。图像深度特征可以在摄像机之间快速在线共享。实验结果表明,该方法的平均检索精度(ARP)为98.74%,模型训练精度为99.51%,帧率为28 FPS。

At present,a large number of offline videos are stored in the servers of the surveillance network in criminal investigation departments of public security organs,and the face retrieval system is designed.The new Quadruplet Network converges faster than the familiar networks such as Alexnet,Googlenet,VGGNet and ResNet.Because of the shared weight design of the network,the retrieval has a high precision,Average Retrieval Precision(ARP)and model training accuracy,and the system has good robustness.The image depth features can be shared quickly online between the cameras.The proposed method is effective,with ARP of 98.74%and a model training accuracy of 99.51%,and a frame rate of 28 FPS.