摘要
提出了多层级特征融合模型,该模型利用深度学习网络提取行人图像的全局特征和局部特征,并将全局和局部特征联合起来,以生成更具辨识度的描述符.在模型中,基于部分的多层级网络用于提取不同网络深度的局部特征,从而将网络底层到高层中提取的局部特征组合起来.全局—局部网络分支则提取网络深层的局部特征和全局特征,用于识别行人.该模型在三个数据集上进行了实验并得到了更好的结果.
This paper proposes a Multi-level Feature Fusion(MFF)model,which uses deep learning networks to extract the global and local features of pedestrian images and combines global and local features to generate more discriminative pedestrian descriptors.In MFF model,Part-based Multi-level Net(PMN)is used to extract local features of different depths of network and combine local features extracted from shallow to deep layers of the network,while Global-Local Branches(GLB)extract the local and global features at the highest level of the network to identify pedestrians.This model has been tested on three widely-used datasets and obtained better results.
作者
吴绍君
高玲
李强
Wu Shaojun;Gao Ling;Li Qiang(School of Information Science and Engineering, Shandong Normal University, 250358, Jinan, China)
出处
《山东师范大学学报(自然科学版)》
CAS
2020年第2期208-216,共9页
Journal of Shandong Normal University(Natural Science)
基金
国家自然科学基金资助项目(61672329).
关键词
图像识别
深度学习
行人重识别
image identification
deep learning
person re-identification