Video Shot Boundary Detection in MPEG Compressed Sequences Using SVM Learning

A number of automated video shot boundary detection methods for indexing a videosequence to facilitate browsing and retrieval have been proposed in recent years.Among these methods,the dissolve shot boundary isn't accurately detected because it involves the camera operation and objectmovement.In this paper,a method based on support vector machine (SVM) is proposed to detect thedissolve shot boundary in MPEG compressed sequence.The problem of detection between the dissolveshot boundary and other boundaries is considered as two-class classification in our method.Featuresfrom the compressed sequences are directly extracted without decoding them,and the optimal classboundary between two classes are learned from training data by using SVM.Experiments,whichcompare various classification methods,show that using proposed method encourages performance ofvideo shot boundary detection.

Rapid Texture Mapping from Image Sequences for Building Geometry Model

An effective approach, mapping the texture for building model based on the digital photogrammetric theory, is proposed. The easily-acquired image sequences from digital video camera on helicopter are used as texture resource, and the correspondence between the space edge in building geometry model and its line feature in image sequences is determined semi-automatically. The experimental results in production of three-dimensional data for car navigation show us an attractive future both in efficiency and effect.

Video expression recognition based on frame-level attention mechanism

Facial expression recognition(FER) in video has attracted the increasing interest and many approaches have been made.The crucial problem of classifying a given video sequence into several basic emotions is how to fuse facial features of individual frames.In this paper, a frame-level attention module is integrated into an improved VGG-based frame work and a lightweight facial expression recognition method is proposed.The proposed network takes a sub video cut from an experimental video sequence as its input and generates a fixed-dimension representation.The VGG-based network with an enhanced branch embeds face images into feature vectors.The frame-level attention module learns weights which are used to adaptively aggregate the feature vectors to form a single discriminative video representation.Finally, a regression module outputs the classification results.The experimental results on CK+and AFEW databases show that the recognition rates of the proposed method can achieve the state-of-the-art performance.

基于改进神经网络的视频序列运动目标识别方法

为改善运动目标检测效果,降低目标漏检率,提出一种基于改进神经网络的视频序列运动目标识别方法。构建改进YOLOv3的运动目标识别模型,以不同帧视频图像为模型输入,经过卷积层的初步特征提取后,输入到由5个残差模块组成的深层网络中。通过以上采样方式构建特征金字塔,实现对运动目标四尺度特征的捕捉。在特征金字塔的每一层,应用K-means算法对运动目标真实框进行聚类,确保候选框尺寸和比例与真实运动目标的统计特性相匹配;再利用获得的候选框和分类器对特征图上每个位置进行目标检测,运用非极大值抑制技术剔除重叠框,将斥力损失函数引入到网络训练总损失之中,使预测框无限贴近运动目标真实框,实现对运动目标的精准识别。实验结果表明,所提方法具有显著的运动目标识别能力,当聚类数目为12时,运动目标识别的AUC、F1指标可达到0.92、0.90,且计算量较少。

基于跟踪检测时序特征融合的视频遮挡目标分割方法

视频实例分割是近年来兴起的一项在图像实例分割基础上引入时序特性的视觉任务,旨在同时对每一帧的目标进行分割并实现帧间的目标跟踪。移动互联网和人工智能的迅猛发展产生了大量的视频数据,但由于拍摄角度、快速运动和部分遮挡等,视频中的物体往往会出现分裂或模糊的情况,使得从视频数据中准确地分割目标并对目标进行处理和分析面临着重大挑战。经查阅和实践发现,现有的视频实例分割方法在遮挡情况下的表现较差。针对上述问题,提出了一种改进的遮挡视频实例分割算法——通过融合Transformer和跟踪检测的时序特征来改善分割性能。为增强网络对空间位置信息的学习能力,该算法将时间维度引入Transformer网络中,并考虑到视频中目标检测、跟踪和分割之间的相互依赖和促进关系,提出了一种能够有效地聚合目标在视频中的跟踪偏移的融合跟踪模块和检测时序特征模块,提升了遮挡环境下的目标分割性能。通过在OVIS和YouTube-VIS数据集上进行的实验,验证了所提方法的有效性。相比当前的基准方法,该方法展现出了更好的分割精度,进一步证明了其优越性。

Video Based Fire Detection Systems on Forest and Wildland Using Convolutional Neural Network

The devastating effects of wildland fire are an unsolved problem,resulting in human losses and the destruction of natural and economic resources.Convolutional neural network(CNN)is shown to perform very well in the area of object classification.This network has the ability to perform feature extraction and classification within the same architecture.In this paper,we propose a CNN for identifying fire in videos.A deep domain based method for video fire detection is proposed to extract a powerful feature representation of fire.Testing on real video sequences,the proposed approach achieves better classification performance as some of relevant conventional video based fire detection methods and indicates that using CNN to detect fire in videos is efficient.To balance the efficiency and accuracy,the model is fine-tuned considering the nature of the target problem and fire data.Experimental results on benchmark fire datasets reveal the effectiveness of the proposed framework and validate its suitability for fire detection in closed-circuit television surveillance systems compared to state-of-the-art methods.

Distortion Identification Technique Based on Hilbert-Huang Transform in Video Stabilization

A distortion identification technique is presented based on Hilbert-Huang transform to identify distortion model and distortion frequency of distorted real-world image sequences. The distortion model is identified simply based on Hilbert marginal spectral analysis after empirical mode decomposing. And distortion frequency is identified by analyzing the occurrence frequency of instantaneous frequency components of every intrinsic mode functions. Rational digital frequency filter with suitable cutoff frequency is designed to remove undesired fluctuations based on identification results. Experimental results show that this technique can identify distortion model and distortion frequency of displacement sequence accurately and efficiently. Based on identification results, distorted image sequence can be stabilized effectively.

Modeling Digital Video Database System with UML

The main purpose of the model is to present how the Unified Modeling Language (UML) can be used for modeling digital video database system (VDBS). It demonstrates the modeling process that can be followed during the analysis phase of complex applications. In order to guarantee the continuity mapping of the models, the authors propose some suggestions to transform the use case diagrams into an object diagram, which is one of the main diagrams for the next development phases.

AUTOMATIC SEGMENTATION OF VIDEO OBJECT PLANES IN MPEG-4 BASED ON SPATIO-TEMPORAL INFORMATION

Segmentation of semantic Video Object Planes (VOP's) from video sequence is a key to the standard MPEG-4 with content-based video coding. In this paper, the approach of automatic Segmentation of VOP's Based on Spatio-Temporal Information (SBSTI) is proposed.The proceeding results demonstrate the good performance of the algorithm.

Similar Video Retrieval via Order-Aware Exemplars and Alignment

In this paper, we present machine learning algorithms and systems for similar video retrieval. Here, the query is itself a video. For the similarity measurement, exemplars, or representative frames in each video, are extracted by unsupervised learning. For this learning, we chose the order-aware competitive learning. After obtaining a set of exemplars for each video, the similarity is computed. Because the numbers and positions of the exemplars are different in each video, we use a similarity computing method called M-distance, which generalizes existing global and local alignment methods using followers to the exemplars. To represent each frame in the video, this paper emphasizes the Frame Signature of the ISO/IEC standard so that the total system, along with its graphical user interface, becomes practical. Experiments on the detection of inserted plagiaristic scenes showed excellent precision-recall curves, with precision values very close to 1. Thus, the proposed system can work as a plagiarism detector for videos. In addition, this method can be regarded as the structuring of unstructured data via numerical labeling by exemplars. Finally, further sophistication of this labeling is discussed.

Research on the Video Detection of Human Respiratory Motion Based on Sequential Images

Objective:Explore the feasibility of the high precision accelerometer for measuring the human respiratory displacement.Methods:A wireless acceleration acquisition system with the low power consumption and the high precision was designed with the high precision acceleration sensor ADXL355 as the core device.Based on the frequency characteristics of the breathing motion and the principle that the displacement can be calculated by the acceleration quadratic integration,two displacement measurement algorithms for the quasi-periodic weak motion are designed.Results:The simulation results show that the proposed algorithm is effective.The experimental results show that the designed acquisition system and algorithm can calculate the human respiratory displacement.Conclusion:The high precision accelerometer can be used to measure the human respiratory displacement,which provides a new method for the measurement of the human respiratory displacement.

一种基于Involution卷积的三维人体重建方法

针对从视频中恢复三维人体模型运动序列时,由于图像特征提取能力有限而导致三维人体模型运动序列重建效果不佳的问题,提出了一种基于Involution卷积的三维人体重建方法。首先为了引入自注意力机制,在ResNet50网络结构中加入Involution算子,获取视频图像帧的特征向量,然后使用姿态估计网络和形状估计网络获取人体姿势以及形状参数,最后使用蒙皮多人线性模型(skinned multi-person linear model, SMPL)生成三维人体模型的运动序列。在三维姿态户外数据集(3D pose in the wild, 3DPW)上与视频人体姿态形状估计推理(video inference for body pose and shape estimation, VIBE)方法以及时间一致性网格恢复(temporally consistent mesh recovery, TCMR)方法进行对比实验,平均精度相比于VIBE、TCMR分别提升了3.1%、0.7%,能够为运动捕捉、三维人体动画制作等工作提供更为准确的三维人体模型。

基于多尺度时空特征的篮球场景中人体姿态估计

人体姿态估计是计算机视觉领域的研究热点之一,目标是从给定的图像或视频中识别人体的关节.由于篮球比赛视频中人体动作复杂多变,易产生运动模糊、遮挡等问题,导致现有的人体姿态算法对篮球动作姿态估计的准确率较低.针对这一问题,提出了一种基于多尺度时空关联特征的篮球动作姿态估计算法,构建基于Transformer的人体时序特征捕捉模块对序列层级的时空特征信息进行建模,以缓解运动模糊、遮挡等现象带来的负面影响.此外,针对人体外形复杂多变的问题,提出了基于可形变卷积的人体空间特征残差融合模块来获取更为充分的空间特征.与现有算法相比,该算法在自行构建的篮球场景人体运动数据集、姿态估计公开基准数据集PoseTrack2017和PoseTrack2018均取得较好的效果.

结合视差补偿与3D数据处理的盲光场图像质量评价

与传统的2D图像相比,光场图像记录了场景中光线的强度和方向信息,在多媒体技术应用领域中占据着重要的地位。但在光场图像的产生、传输等处理过程中会不可避免地引入失真,影响用户视觉体验,因而需构建有效、准确的光场图像质量评价方法对其进行评估。本文基于光场图像的伪视频和极平面图像提出了结合视差补偿与3D数据处理的盲光场图像质量评价方法。首先,采用视差补偿模块处理光场图像伪视频序列,然后对经过视差补偿后的伪视频序列进行三维离散小波变换以及三维去均值对比度归一化处理,并提取频域和空域特征。同时,在极平面图像上提取方向梯度直方图特征表征其角度信息的失真。最后,运用支持向量机建立图像质量评价指标到主观质量得分的回归模型。所提算法在公开数据库NBU-LF1.0,Win5-LID和SHU上的PLCC分别达到0.886 1,0.928 7和0.976 9,实验结果表明,与经典的二维图像质量评价方法以及先进的光场图像质量评价方法相比,本文方法与主观质量评价结果的一致性更高。

视频片段教学法在儿童口腔科进修生临床前培训中的应用

目的 探究在传统讲座教学和视频片段教学模式下儿童口腔科进修生对于儿童患者行为管理技巧的掌握情况。方法 在儿童口腔科临床前培训课程开始前,59名进修生被随机分为两组。传统讲座教学组29人,视频片段教学组30人。每组均向学员讲授六种儿童口腔行为管理技巧。在课程结束后,学员填写调查问卷。通过调查问卷评估学员对六种行为管理技巧的认知和临床应用的意愿。在其临床实践结束后,同样就行为管理技巧应用情况完成调查问卷。结果 课程结束后,更多学员倾向于将视频片段教学模式用于儿童口腔医学行为管理技巧的教学。对于声音控制,分散注意力和非语言交流三种行为管理技巧,视频片段教学组的学员在课后认为其临床应用较为简易的人数显著多于传统讲座教学组的学员(P<0.05)。且视频片段教学能够显著提高学员对告知-示范-操作、声音控制、分散注意力和非语言交流这四种行为管理技巧临床应用的意愿(P<0.05),同时显著提高学员在临床实践中采用告知-示范-操作,声音控制和分散注意力这三种行为管理技巧的比率(P<0.05)。结论 在儿童口腔医学行为管理技巧的授课中,视频片段教学方式能够使学员更好地掌握相关技巧并且提高其临床应用的意愿。这一教学模式具有较好的应用和推广前景。

快速顺序诱导方法的研究新进展

快速顺序诱导(rapid sequence induction,RSI)是指应用快速起效的镇静药、镇痛药及肌松剂使患者短时间内意识及自主呼吸消失,肌肉松弛,达到满意的气管插管条件。随着临床理论研究与技术的进步,经典RSI的操作方法、诱导药物组成和选择等进行了不同程度的更新与发展。尤其在新型冠状病毒疫情爆发期间,新型冠状肺炎患者的病理生理特殊性,救治这些患者所实施的快诱导插管成为了临床医师的又一挑战。因此,本文将对RSI主要的争议点及最新的临床研究进行综述,以期能为临床工作提供一定的参考价值。

基于深度学习的行人再识别技术分析

阐述在视频监控系统的辅助识别中,行人再识别技术能够有效提升视频监控系统的效率与质量,探讨以深度学习为基础的行人再识别技术特点、再识别技术的表征研究法、度量研究法、局部表征研究法、视频序列研究法的分析。

基于递归长短期记忆网络和镜头序列注意网络的视频摘要生成

为解决基于长短期记忆网络LSTM的视频摘要生成方法当输入序列过长时LSTM网络中的记忆单元不能集中在长时间序列的跨度上。通过深度学习的方法研究了一种基于递归长短期记忆网络(ReLSTM)和序列注意(sequential attention,SSA)的视频摘要生成模型用以提高深度学习网络学习时序特征的能力。该模型使用ReLSTM网络提取时间特征。同时,利用SSA动态调整每个视频序列输入到ReLSTM网络中的特征权重。结果表明:在数据集TVSum上F 1-score平均提高2.5%,最高提高0.2%。在数据集SumMe上F 1-score平均提高7.8%,最高提高3.4%。可见本文方法能有效地学习镜头之间的时序特征。

基于混合注意力机制的视频序列表情识别

针对视频序列表情识别方法中模型庞大、提取到的特征表征能力低等困难,提出了一种基于混合注意力机制的视频序列表情识别方法。利用改进的卷积神经网络与门控循环单元提取表情序列的时空信息。通过通道注意力和时间注意力组成的混合注意力模块,有效地增强对特征通道的敏感性,同时能够聚焦于表情峰值帧,提取最具有表达能力的表情特征。提出模型在CK+与Oulu-CASIA两个公共数据集分类准确率各达到了98.70%、89.15%,与其它主流算法对比,该方法具有更好的识别性能。

基于深度学习的体育视频关键姿态提取研究

针对体育视频中运动目标的关键姿态不一,人工提取视频图像中的关键姿态费时费力的问题,提出一种基于深度学习的关键姿态提取算法。利用距离变换根据体育视频中像素点特征,对关键姿态进行判断,分别运用骨架候选点和背景候选点,计算体育视频中运动目标的关键姿态。实验表明,提出的算法能够准确清晰提取视频中的运动目标特征轮廓,并能够区分关键姿态点和轮廓点,从而实现体育视频的人体关键姿态的提取。