面向轨道分割与侵入物检测的多任务学习

轨道车智能防护会涉及轨道车侵入物检测与行驶区域分割任务,在深度学习领域已有针对各任务的算法,却无法很好满足多任务情形时的需求.该算法使用轻量级卷积神经网络(CNN)作为编码器提取特征图,随之将特征图送到两个基于one-stage检测网络的解码器中,进而完成各自的任务.不同级别和尺度的语义特征在编码器输出的特征图中被融合,良好地完成像素级语义预测,在检测和分割效果上有明显提升.采用本算法的设备将掌握对新目标的识别检测判断与追踪,为提升轨道车行驶安全做出保障.     

深度卷积神经网络图像实例分割方法研究进展

图像实例分割是图像处理和计算机视觉技术中关于图像理解的重要环节,随着深度学习和深层卷积神经网络日趋成熟,基于深度卷积神经网络的图像实例分割方法取得了跨越性进展.实例分割任务实际上是目标检测和语义分割两项任务的结合,可以在像素层面完成识别图像中目标轮廓的任务.实例分割不仅可以定位图像中目标的位置,从像素层面上分割所有目标...     

基于形式概念分析和语义关联规则的目标图像标注

基于目标的图像标注一直是图像处理和计算机视觉领域中一个重要的研究问题.图像目标的多尺度性、多形变性使得图像标注十分困难.目标分割和目标识别是目标图像标注任务中两大关键问题.本文提出一种基于形式概念分析(Formal concept analysis, FCA)和语义关联规则的目标图像标注方法, 针对目标建议算法生成图像块中存在的高度重叠问题, 借鉴形式概念分析中概念格的思想, 按照图像块的共性将其归成几个图像簇挖掘图像类别模式, 利用类别概率分布判决和平坦度判决分别去除目标噪声块和背景噪声块, 最终得到目标语义簇; 针对语义目标判别问题, 首先对有效图像簇进行特征融合形成共性特征描述, 通过分类器进行类别判决, 生成初始目标图像标注, 然后利用图像语义标注词挖掘语义关联规则, 进行图像标注的语义补充, 以避免挖掘类别模式时丢失较小的语义目标.实验表明, 本文提出的图像标注算法既能保证语义标注的准确性, 又能保证语义标注的完整性, 具有较好的图像标注性能.     

基于去冗余特征和语义关系约束的零样本属性识别

基于生成式的零样本识别方法在生成特征时受冗余信息和域偏移的影响,识别精度不佳.针对此问题,文中提出基于去冗余特征和语义关系约束的零样本属性识别方法.首先,将视觉特征映射到一个新的特征空间,通过互相关信息对视觉特征进行去冗余处理,在去除冗余视觉特征的同时保留类别的相关性,由于在识别过程中减少冗余信息的干扰,从而提高零样本识别的精度.然后,利用可见类和不可见类之间的语义关系建立知识迁移模型,并引入语义关系约束损失,约束知识迁移的过程,使生成器生成的视觉特征更能反映可见类和不可见类之间语义关系,缓解两者之间的域偏移问题.最后,引入循环一致性结构,使生成的伪特征更接近真实特征.在数据集上的实验证实文中方法提高零样本识别任务的精度,并具有较优的泛化性能.     

结合环状原型空间优化的开放集目标检测

目的 现有目标检测任务常在封闭集设定中进行。然而在现实问题中,待检测图片中往往包含未知类别目标。因此,在保证模型对已知类检测性能的基础上,为了提升模型在现实检测任务中对新增类别的目标检测能力,本文对开放集目标检测任务进行研究。方法 区别于现有的开放集目标检测框架在检测任务中将背景类与未知类视为一个类别进行优化,本文框架在进行开放集类别识别的过程中,优先识别候选框属于背景类或是含待识别目标类别,而后再对含待识别目标类别进行已知类与未知类的判别。本文提出基于环状原型空间优化的检测器,该检测器可以通过优化待检测框的特征在高维空间中的稀疏程度对已知类、未知类与背景类进行环状序列判别,从而提升模型对开放集类别的检测性能。在（region proposal networks,RPN）层后设计了随机覆盖候选框的方式筛选相关的背景类训练框,避免了以往开放集检测工作中繁杂的背景类采样步骤。结果 本文方法在保证模型对封闭集设定下检测性能的情况下,通过逐步增加未知类别的数量,在Visual Object Classes-Common Objects in Context-20(VOC-COCO-20),Vi...     

目标语义概率模型在类目标识别和地物场景分析中的算法研究

基于文本分析统计模型提出了图像类目标的语义概率模型,并且将这种概率模型应用于目标识别和复杂场景下的地物分析.首先将图像表示成多个特征局部区域的集合,然后根据目标语义概率模型得到图像、特征局部和目标语义之间的概率关系,通过计算后验概率可以实现目标语义类别的识别.目标概率模型通过EM算法获得模型估计参数.实验结果显示,在识别复杂背景中的目标达到了很好的效果.场景分析中根据图像中各局部区域与目标语义的概率分布可以实现场景中感兴趣区域的标注,实验结果说明此方法有可行性.     

一种新的强分辨性零样本目标检测方法

零样本目标检测是近年来用于对训练中未见（unseen）类别目标进行分类和定位的一种技术。由此带来了目标检测中的新问题——目标视觉特征与其对应的类别语义信息映射关系不准确,未见（unseen）类别目标与背景分辨性不强。提出的强分辨算法（adaptive channel with center distance Gaussian distribution loss,ACDG）使用特征图通道自适应加权机制,能够根据不同类别目标对特征图通道自动赋予权值,加强关键特征图的权重,抑制次要特征图的信息,增强特征提取网络的表征性,以建立更准确的视觉特征和语义特征之间映射关系。为了解决未见类别目标与背景分辨性不高的问题,提出中心距离高斯分布损失,约束预测边界框中心点与真实值（groundtruth）中心点位置的距离,进而加快损失函数收敛。为了验证所提算法的先进性,在MS COCO数据集上完成了大量实验,召回率和平均精度分别高出原始方法5.9%和4.5%。     

自适应上下文特征的多尺度目标检测算法

识别多尺度目标是检测任务中的一项挑战,针对检测中的多尺度问题,提出自适应上下文特征的多尺度目标检测算法。针对不同尺度的目标需要不同大小感受野特征进行识别的问题,构建了一种多感受野特征提取网络,通过多分支并行空洞卷积,从高层语义特征中挖掘标签中的上下文信息;针对不同尺度目标的语义特征出现在不同分辨率特征图中的问题,基于改进的通道注意力机制,提出自适应的特征融合网络,通过学习不同分辨率特征图之间的相关性,在全局语义特征中融合局部位置特征;利用不同尺度的特征图识别不同尺度的物体。在PASCAL VOC数据集上对本文算法进行验证,本文方法的检测精度达到了85.74%,相较于Faster R-CNN检测精度提升约8.7%,相较于基线检测算法YOLOv3+提升约2.06%。     

基于数据合成和度量学习的台标检测与识别

台标是视频的重要语义信息,其检测与识别面临类别多、结构复杂、区域小、信息量低、背景干扰大等难题.为提高模型的泛化能力,本文提出将台标图像叠加到背景图像中合成台标数据,来构建训练数据集.进一步,我们提出两阶段可伸缩台标检测与识别（scalable logo detection and recognition,SLDR）方法,其采用batch-hard度量学习方法快速训练匹配模型,确定台标类别.SLDR的检测与识别分离机制使得其可将检测目标扩展到未知类别.实验结果表明,合成数据可以有效提升模型的泛化能力和检测精度.实验亦显示SLDR方法在不更新检测模型的情况下,即可获得与端到端模型相当的精度.     

任务感知双原型网络的人物交互少样本识别

人物交互(HOI)识别是计算机视觉领域的重要研究热点.随着深度学习在图像分类任务中的巨大成功,人物交互识别任务也取得重大进展,但样本不平衡和组合爆炸问题仍是制约当前人物交互识别任务性能的关键挑战.由此,将人物交互识别任务与少样本学习相结合,将人物交互识别任务定义为一个少样本任务,并提出了任务感知双原型网络(TDP-Net)来解决少样本人物交互任务.具体地,首先使用图方法为每个任务生成语义感知的任务表示作为任务的先验信息,并使用语义图注意力模块(SGA-Module)生成注意力权重,对特征图中不同区域进行不同重要程度的关注,以适应不同任务条件下的映射关系,实现在新任务中自动推理.此外,还设计了一个双路原型模块(DP-Module)以分别产生交互类别的动作类原型和物体类原型,并分别对动词和名词进行分类.通过分别为动作和物体建立类原型,有效地分离了动作和物体间复杂的视觉关系.同时由于人物交互类别之间具有相似性,可通过重新组合动作和物体类别将知识迁移到新的交互类别中.实验结果表明,该模型在人物交互少样本任务上的平均准确率比基线方法在两个实验设置上分别提高了3.2个百分点和15.7个百分点,验证了TDP-Net在少样本人物交互任务中的有效性.     

A novel unsupervised 3D skeleton detection in RGB-D images for video surveillance

In this paper we present a novel moment-based skeleton detection for representing human objects in RGB-D videos with animated 3D skeletons. An object often consists of several parts, where each of them can be concisely represented with a skeleton. However, it remains as a challenge to detect the skeletons of individual objects in an image since it requires an effective part detector and a part merging algorithm to group parts into objects. In this paper, we present a novel fully unsupervised learning framework to detect the skeletons of human objects in a RGB-D video. The skeleton modeling algorithm uses a pipeline architecture which consists of a series of cascaded operations, i.e., symmetry patch detection, linear time search of symmetry patch pairs, part and symmetry detection, symmetry graph partitioning, and object segmentation. The properties of geometric moment-based functions for embedding symmetry features into centers of symmetry patches are also investigated in detail. As compared with the state-of-the-art deep learning approaches for skeleton detection, the proposed approach does not require tedious human labeling work on training images to locate the skeleton pixels and their associated scale information. Although our algorithm can detect parts and objects simultaneously, a pre-learned convolution neural network (CNN) can be used to locate the human object from each frame of the input video RGB-D video in order to achieve the goal of constructing real-time applications. This much reduces the complexity to detect the skeleton structure of individual human objects with our proposed method. Using the segmented human object skeleton model, a video surveillance application is constructed to verify the effectiveness of the approach. Experimental results show that the proposed method gives good performance in terms of detection and recognition using publicly available datasets.

     

一种有效的运动物体检测方法

为了从监控视频中检测出较高质量的运动物体,文章提出了一种基于帧间差分和背景差分相结合的运动目标的检测方法,并且采用像素级和帧级背景更新相配合的一种背景更新策略。算法求取各像素点处的最大概率灰度,从而提取出连续视频的背景图像;相邻帧则利用帧间差分法以及背景差分法得到两幅运动区域图像;将两幅运动区域图像相与,提取出较为准确的运动目标。实验证明,该算法对光线的变化鲁棒性较高,运算速度较快,且能够及时的响应监控视频的实时变化,提高运动目标的检测质量。     

基于帧间差分法的动体特征速度聚类分析

针对智能视频监控中快速、准确的检测和识别运动物体的问题,提出了一种依据运动物体特征速度来检测识别动体以及解读其语义含义的算法。该方法以相对帧间差分法为基础,通过对预处理后的二值斑块图像的标记,计算斑块的像素长度作为其特征速度,并依据斑块特征速度的众数进行聚类分析,从斑块特征速度得到运动物体的特征速度语义解读和运动物体的检测识别。实验结果表明,斑块的特征速度不仅可以实现对运动物体的检测,而且通过聚类分析可以准确的得出动体特征的语义解读。用特征速度和众数聚类分析方法实现对运动物体的检测识别和语义解读相对于其他统计算法简单有效,便于智能摄像机的嵌入式开发。     

基于帧间差分和边缘差分的遗留物检测算法

遗留物检测是智能视频监控系统的核心功能,遗留物一般较小,所处环境复杂,传统的运动目标检测算法直接用于遗留物检测效果一般.提出了一种基于帧间差分与边缘差分的遗留物检测算法,首先进行帧间差分得到运动目标区域,然后将当前帧图像和前一帧的背景图像进行边缘差分运算得到运动目标的边缘,融合二次差分的结果即可得到运动目标的完整轮廓特征,最终通过判断运动目标在场景中的滞留时间是否达到或超过报警系统设置的阈值来标示遗留物,供智能视频监控系统处理.实验结果证明该算法实时性好且识别率较高.     

Multi-object trajectory tracking

The majority of existing tracking algorithms are based on the maximum a posteriori solution of a probabilistic framework using a Hidden Markov Model, where the distribution of the object state at the current time instance is estimated based on current and previous observations. However, this approach is prone to errors caused by distractions such as occlusions, background clutters and multi-object confusions. In this paper, we propose a multiple object tracking algorithm that seeks the optimal state sequence that maximizes the joint multi-object state-observation probability. We call this algorithm trajectory tracking since it estimates the state sequence or “trajectory” instead of the current state. The algorithm is capable of tracking unknown time-varying number of multiple objects. We also introduce a novel observation model which is composed of the original image, the foreground mask given by background subtraction and the object detection map generated by an object detector. The image provides the object appearance information. The foreground mask enables the likelihood computation to consider the multi-object configuration in its entirety. The detection map consists of pixel-wise object detection scores, which drives the tracking algorithm to perform joint inference on both the number of objects and their configurations efficiently. The proposed algorithm has been implemented and tested extensively in a complete CCTV video surveillance system to monitor entries and detect tailgating and piggy-backing violations at access points for over six months. The system achieved 98.3% precision in event classification. The violation detection rate is 90.4% and the detection precision is 85.2%. The results clearly demonstrate the advantages of the proposed detection based trajectory tracking framework.     

一种基于梯度方向信息的运动目标检测算法

运动目标检测是智能视觉监控系统的基本内容。在对现有算法分析的基础上提出了一种基于梯度方向信息的运动目标检测算法。首先利用方向信息提取视频图像序列中每一帧的边缘梯度图,然后通过改进传统帧差算法,采用uint8数据格式处理含有时间关系的两帧图像以此确定运动目标粗略边界,经运动目标连通域识别,最后结合梯度方向信息准确提取运动目标的完整轮廓。实验结果表明,该算法克服了传统帧差算法不能准确定位目标的缺点,在室内外复杂背景下均能准确地提取完整的目标轮廓。     

Multiresolution video object extraction fitted to scalable wavelet-based object coding

To enable content based functionalities in video processing algorithms, decomposition of scenes into semantic objects is necessary. A semi-automatic Markov random field based multiresolution algorithm is presented for video object extraction in a complex scene. In the first frame, spatial segmentation and user intervention determine objects of interest. The specified objects are subsequently tracked in successive frames and newly appeared objects/regions are also detected. The video object extraction algorithm includes discrete wavelet transform decomposition multiresolution Markov random field (MRF)-based spatial segmentation with emphasis on border smoothness at different resolutions, and an MRF-based backward region classification that determines the tracked objects in the scene. Finally, a motion constraint, embedded in the region classifier, determines the newly appeared objects/regions and completes the proposed algorithm towards an efficient video segmentation algorithm. The results are applicable for generic segmentation applications, however the proposed multiresolution video segmentation algorithm supports scalable object-based wavelet coding in particular. Moreover, compared to traditional object extraction algorithms, it produces smoother and more visually pleasing shape masks at different resolutions. The proposed effective multiresolution video object extraction method allows for larger motion, better noise tolerance and less computational complexity     

A novel framework for intelligent surveillance system based on abnormal human activity detection in academic environments

Abnormal activity detection plays a crucial role in surveillance applications, and a surveillance system that can perform robustly in an academic environment has become an urgent need. In this paper, we propose a novel framework for an automatic real-time video-based surveillance system which can simultaneously perform the tracking, semantic scene learning, and abnormality detection in an academic environment. To develop our system, we have divided the work into three phases: preprocessing phase, abnormal human activity detection phase, and content-based image retrieval phase. For motion object detection, we used the temporal-differencing algorithm and then located the motions region using the Gaussian function. Furthermore, the shape model based on OMEGA equation was used as a filter for the detected objects (i.e., human and non-human). For object activities analysis, we evaluated and analyzed the human activities of the detected objects. We classified the human activities into two groups: normal activities and abnormal activities based on the support vector machine. The machine then provides an automatic warning in case of abnormal human activities. It also embeds a method to retrieve the detected object from the database for object recognition and identification using content-based image retrieval. Finally, a software-based simulation using MATLAB was performed and the results of the conducted experiments showed an excellent surveillance system that can simultaneously perform the tracking, semantic scene learning, and abnormality detection in an academic environment with no human intervention.

     

基于实例分割模型优化的道路抛洒物检测算法

在交通安全领域,道路抛洒物易引发交通事故,构成了交通安全隐患。针对传统抛洒物检测方式识别率低、对于多类抛洒物检测效果不佳等问题,提出了一种基于实例分割模型CenterMask优化的道路抛洒物检测算法。首先,使用空洞卷积优化的残差网络ResNet50作为主干神经网络来提取特征并进行多尺度处理;然后,通过距离交并比（DIoU）函数优化的全卷积单阶段（FCOS）目标检测器实现对抛洒物的检测和分类;最后,使用空间注意力引导掩膜作为掩膜分割分支来实现对于目标形态的分割,并采用迁移学习的方式实现模型的训练。实验结果表明,所提算法对于抛洒物目标的检测率为94.82%,相较常见实例分割算法Mask R-CNN,所提的道路抛洒物检测算法在边界框检测上的平均精度（AP）提高了8.10个百分点。     

视频监控中一种完整提取运动目标的检测算法

提出一种视频监控中完整、精确提取运动目标前景的检测算法.首先对彩色图像建立混合高斯模型,由背景差分法得到基本准确的前景图像;然后和对称差分法图像综合,得到完整可靠的运动目标图像;再利用亮度信息消除运动目标阴影;最后利用形态学滤波和连通区域面积检测进行后处理.实验结果表明,该算法检测的运动目标前景信息完整准确,对固定场景下的视频监控系统具有一定实用价值.