夜间红外行人图像在可见光行人图像库的检索算法  

Retrieval Algorithm on Checking Night-shot Infrared Pedestrian’s Image into Visible-light-rendering Image Library of Pedestrians

作　　者：李宗辰 叶东 李俊瑶 张颐 陈春涛[1] LI Zongchen;YE Dong;LI Junyao;ZHANG Yi;CHEN Chuntao(Department of Forensic Science,Jiangsu Police Institute,Nanjing 210031,China)

机构地区：[1]江苏警官学院刑事科学技术系,南京210031

出　　处：《刑事技术》2023年第4期378-385,共8页Forensic Science and Technology

基　　金：江苏省公安厅科技研究项目(2020KX006);江苏省高等学校自然科学面上项目(21KJD510005)。

摘　　要：本文旨在弥补夜间红外模式监控下行人图像与白天监控下行人图像间的模态差异和异构特性,较好实现多摄像头下红外人像与可见光人像数据间的检索和比对任务。设计一种局部参数共享的双路卷积神经网络模型,提取具有全局粗粒度和局部细粒度的两种人像特征,通过对两种特征分别进行特征度量,实现跨模态人像检索模型的训练和优化。结果表明,模型所提取的两种特征在测试数据集上具有较好的应用效果,准确率与目前主流算法相比具有竞争力。本研究能够服务视频侦查工作,提高现有的动态、静态人像比对的应用水平。Image retrieval algorithms aim at achieving individual re-identifi cation through searching across a gallery of people’s images from different non-overlapping video monitoring cameras.However,night-shot infrared pedestrians’images captured by surveillance systems have different data distributions from those of conventional colored RGB images,resulting in loss of all the color components which are crucial for pedestrian comparison.The differences between the two kinds of images are regarded as their modality discrepancy that can lead to large intra-class variations and modality gaps across different cameras.Thus,it is necessary and valuable to effectively bridge the modality gaps and alleviate the heterogeneous characteristics between pedestrian images shot nightly in infrared mode and those monitor-kept in the daytime so that the available retrieval algorithms could be improved of their effects under the widely-used visible-light-rendering image database.Accordingly,a dual-path partial-parameter-sharing framework was here designed with combination into a deep convolutional neural network(CNN)to extract the features from two�modality images possessing both the global coarse-granularity and local fi ne-granularity.The pre-trained ResNet50-based network was used to optimize the model training with the constraints of cross-modality entropy loss,center-based and sample-based triplet loss,eventually having brought forth an optimal model through adjustment of the weighted constraints into functions of the three losses indicated above.Subsequently,the feature measurement was carried out to validate the recognition effect on cross-modality images.The experimental results showed that the extracted multi-granularity features presented good application effect on the tested dataset under single shot mode,having SYSU-MM01 dataset demonstrated of effect verifi cation and visualization that the resulting multi-granularity components did complement each other and hold a competitive testing accuracy compared to other several

关 键 词：红外人像 跨模态 行人重识别 多粒度 卷积神经网络 

分 类 号：DF793.2[政治法律—诉讼法学]