俯视深度头肩序列行人再识别  

Person re-identification based on top-view depth head and shoulder sequence

作　　者：王新年[1] 刘春华 齐国清[1] 张世强 Wang Xinnian;Liu Chunhua;Qi Guoqing;Zhang Shiqiang(Dalian Maritime University,Dalian 116026,China;Hualuzhida Technology Co.,Ltd.,Dalian 116023,China)

机构地区：[1]大连海事大学,大连116026 [2]华录智达科技有限公司,大连116023

出　　处：《中国图象图形学报》2020年第7期1393-1407,共15页Journal of Image and Graphics

基　　金：大连市科技创新基金项目(2019J12GX036)。

摘　　要：目的行人再识别是指在一个或者多个相机拍摄的图像或视频中实现行人匹配的技术,广泛用于图像检索、智能安保等领域。按照相机种类和拍摄视角的不同,行人再识别算法可主要分为基于侧视角彩色相机的行人再识别算法和基于俯视角深度相机的行人再识别算法。在侧视角彩色相机场景中,行人身体的大部分表观信息可见;而在俯视角深度相机场景中,仅行人头部和肩部的结构信息可见。现有的多数算法主要针对侧视角彩色相机场景,只有少数算法可以直接应用于俯视角深度相机场景中,尤其是低分辨率场景,如公交车的车载飞行时间(time of flight,TOF)相机拍摄的视频。因此针对俯视角深度相机场景,本文提出了一种基于俯视深度头肩序列的行人再识别算法,以期提高低分辨率场景下的行人再识别精度。方法对俯视深度头肩序列进行头部区域检测和卡尔曼滤波器跟踪,获取行人的头部图像序列,构建头部深度能量图组(head depth energy map group,He DEMaG),并据此提取深度特征、面积特征、投影特征、傅里叶描述子和方向梯度直方图(histogram of oriented gradient,HOG)特征。计算行人之间头部深度能量图组的各特征之间的相似度,再利用经过模型学习所获得的权重系数对各特征相似度进行加权融合,从而得到相似度总分,将最大相似度对应的行人标签作为识别结果,实现行人再识别。结果本文算法在公开的室内单人场景TVPR(top view person re-identification)数据集、自建的室内多人场景TDPI-L(top-view depth based person identification for laboratory scenarios)数据集和公交车实际场景TDPI-B(top-view depth based person identification for bus scenarios)数据集上进行了测试,使用首位匹配率(rank-1)、前5位匹配率(rank-5)、宏F1值(macro-F1)、累计匹配曲线(cumulative match characteristic,CMC)和平均耗时等5个指标来衡量算法性能。�Objective Person reidentification is an important task in video surveillance systems with a goal to establish the correspondence among images or videos of a person taken from different cameras at different times.In accordance with camera types,person re-identification algorithms can be divided into RGB camera-based and depth camera-based ones.RGB camera-based algorithms are generally based on the appearance characteristics of clothes,such as color and texture.Their performances are greatly affected by external conditions,such as illumination variations.On the contrary,depth camerabased algorithms are minimally affected by lighting conditions.Person re-identification algorithms can also be divided into side view-oriented and vertical view-oriented algorithms according to camera-shooting angle.Most body parts can be seen in side-view scenarios,whereas only the plan view of head and shoulders can be seen in vertical-view scenarios.Most existing algorithms are for side-view RGB scenarios,and only a few of them can be directly applied to top-view depth scenarios.For example,they have poor performance in the case of bus-mounted low-resolution depth cameras.Our focus is on person re-identification on depth head and shoulder sequences.Method The proposed person re-identification algorithm consists of four modules,namely,head region detection,head depth energy map group(He DEMaG)construction,He DEMaG-based multifeature representation and similarity computation,and learning-based score-level fusion and person re-identification.First,the head region detection module is to detect each head region in every frame.The pixel value in a depth image represents the distance between an object and the camera plane.The range that the height of a person distributes is used to roughly segment the candidate head regions.A frame-averaging model is proposed to compute the distance between floor and the camera plane for determining the height of each person with respect to floor.The person’s height can be computed by subtracting floor val

关 键 词：深度相机 俯视深度头肩序列 头部深度能量图组 相似度权重学习 行人再识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]