公开(公告)号：US20100266159A1

公开(公告)日：2010-10-21

申请号：US12427353

申请日：2009-04-21

申请人： Kazuya UEKI ,  Yihong Gong ,  Fengjun Lv

发明人： Kazuya UEKI ,  Yihong Gong ,  Fengjun Lv

IPC分类号： G06K9/00

CPC分类号： G06K9/00362 ,  G06K9/00261 ,  G06K9/00295

摘要： A human tracking apparatus and method capable of highly accurately tracking the movement of persons photographed in moving images includes: an image memory 107 that stores an inputted frame image; a human detecting unit 101 that detects persons photographed in the inputted frame image; a candidate registering unit 106 that registers already detected persons as candidates; a similarity index calculating unit 102 that calculates similarity indices indicating the similarity between the persons detected in the inputted frame image and the registered candidates for two or more types of parameters based on the stored frame images in relation to all combinations of the persons and the candidates; a normalizing unit 103 that normalizes the similarity indices; an integrating unit 104 that integrates the normalized indices for each combination of the detected persons and the candidates; and a tracking unit 105 that identifies a person the same as an arbitrary candidate based on the similarity indices.

摘要翻译： 能够高精度地跟踪在运动图像中拍摄的人的移动的人体跟踪装置和方法包括：存储输入的帧图像的图像存储器107; 人物检测单元101，其检测在输入的帧图像中拍摄的人物; 将已经检测到的人登记为候选的候选登记单元106; 相似度计算单元102，其基于与人和候选者的所有组合相关联的所存储的帧图像，计算指示在输入的帧图像中检测到的人与两个或多个类型的参数之间的相似度的相似性 ; 归一化单元103，其使相似性指数标准化; 整合单元104，其将检测到的人和候选的每个组合的归一化指数进行积分; 以及跟踪单元105，其基于相似性索引来识别与任意候选者相同的人。

公开(公告)号：US20090296985A1

公开(公告)日：2009-12-03

申请号：US12277278

申请日：2008-11-24

申请人： Fengjun Lv ,  Wei Xu ,  Yihong Gong

发明人： Fengjun Lv ,  Wei Xu ,  Yihong Gong

IPC分类号： G06K9/00 ,  H04N5/225

CPC分类号： G06T7/292 ,  G06T7/285 ,  G06T2207/10016 ,  G06T2207/30196 ,  G06T2207/30232 ,  G06T2207/30241

摘要： System and methods are disclosed to perform multi-human 3D tracking with a plurality of cameras. At each view, a module receives each camera output and provides 2D human detection candidates. A plurality of 2D tracking modules are connected to the CNNs, each 2D tracking module managing 2D tracking independently. A 3D tracking module is connected to the 2D tracking modules to receive promising 2D tracking hypotheses. The 3D tracking module selects trajectories from the 2D tracking modules to generate 3D tracking hypotheses.

摘要翻译： 公开了用多个摄像机进行多人3D跟踪的系统和方法。 在每个视图中，模块接收每个摄像机输出并提供2D人类检测候选。 多个2D跟踪模块连接到CNN，每个2D跟踪模块独立管理2D跟踪。 3D跟踪模块连接到2D跟踪模块，以接收有希望的2D跟踪假设。 3D跟踪模块从2D跟踪模块中选择轨迹，以生成3D跟踪假设。

公开(公告)号：US20090141969A1

公开(公告)日：2009-06-04

申请号：US12277504

申请日：2008-11-25

申请人： Kai Yu ,  Wei Xu ,  Yihong Gong

发明人： Kai Yu ,  Wei Xu ,  Yihong Gong

IPC分类号： G06K9/62

CPC分类号： G06K9/6256 ,  G06N3/08

摘要： A method and system for training a neural network of a visual recognition computer system, extracts at least one feature of an image or video frame with a feature extractor; approximates the at least one feature of the image or video frame with an auxiliary output provided in the neural network; and measures a feature difference between the extracted at least one feature of the image or video frame and the approximated at least one feature of the image or video frame with an auxiliary error calculator. A joint learner of the method and system adjusts at least one parameter of the neural network to minimize the measured feature difference.

摘要翻译： 一种用于训练视觉识别计算机系统的神经网络的方法和系统，使用特征提取器提取图像或视频帧的至少一个特征; 使用在神经网络中提供的辅助输出近似图像或视频帧的至少一个特征; 并且利用辅助误差计算器测量提取的图像或视频帧的至少一个特征与图像或视频帧的近似的至少一个特征之间的特征差异。 该方法和系统的联合学习者调整神经网络的至少一个参数以最小化测量的特征差异。

公开(公告)号：US20070103595A1

公开(公告)日：2007-05-10

申请号：US11553552

申请日：2006-10-27

申请人： Yihong Gong ,  Mei Han ,  Dan Kong ,  Hai Tao ,  Wei Xu

发明人： Yihong Gong ,  Mei Han ,  Dan Kong ,  Hai Tao ,  Wei Xu

IPC分类号： H04N5/00

CPC分类号： H04N5/262 ,  H04N19/587 ,  H04N19/59

摘要： A video super-resolution method that combines information from different spatial-temporal resolution cameras by constructing a personalized dictionary from a high resolution image of a scene resulting in a domain specific prior that performs better than a general dictionary built from images.

摘要翻译： 一种视频超分辨率方法，其通过从场景的高分辨率图像构造个性化字典来组合来自不同空间 - 时间分辨率相机的信息，从而产生比由图像构建的通用字典更好的域特定。

公开(公告)号：US06925455B2

公开(公告)日：2005-08-02

申请号：US10011215

申请日：2001-10-25

申请人： Yihong Gong ,  Xin Liu

发明人： Yihong Gong ,  Xin Liu

IPC分类号： H04N5/91 ,  G06F17/30 ,  G06K9/62 ,  G06K9/68 ,  G10L15/00 ,  G10L15/06 ,  G10L15/10 ,  G10L15/16 ,  G10L15/26 ,  G10L17/00 ,  G10L17/26 ,  H04N7/16 ,  H04N21/2368 ,  H04N21/266 ,  H04N21/434 ,  G06N5/00

CPC分类号： G06K9/6293 ,  G06F17/30029 ,  G06F17/30056 ,  G06F17/30787 ,  G06F17/30796 ,  G06F17/30843 ,  G10L15/26 ,  H04N7/165 ,  H04N21/2368 ,  H04N21/26603 ,  H04N21/4341

摘要： Systems and methods create high quality audio-centric, image-centric, and integrated audio-visual summaries by seamlessly integrating image, audio, and text features extracted from input video. Integrated summarization may be employed when strict synchronization of audio and image content is not required. Video programming which requires synchronization of the audio content and the image content may be summarized using either an audio-centric or an image-centric approach. Both a machine learning-based approach and an alternative, heuristics-based approach are disclosed. Numerous probabilistic methods may be employed with the machine learning-based learning approach, such as naïve Bayes, decision tree, neural networks, and maximum entropy. To create an integrated audio-visual summary using the alternative, heuristics-based approach, a maximum-bipartite-matching approach is disclosed by way of example.

摘要翻译： 系统和方法通过无缝集成从输入视频中提取的图像，音频和文本功能，创建高质量的以音频为中心，以图像为中心的集成视听摘要。 当不需要音频和图像内容的严格同步时，可以采用集成摘要。 可以使用以音频为中心或以图像为中心的方法来总结需要音频内容和图像内容的同步的视频节目。 公开了基于机器学习的方法和基于启发式的替代方法。 可以采用基于机器学习的学习方法的许多概率方法，例如朴素贝叶斯，决策树，神经网络和最大熵。 为了使用替代的基于启发式的方法来创建集成的视听摘要，通过示例的方式公开了最大二分法匹配方法。

公开(公告)号：US20050105764A1

公开(公告)日：2005-05-19

申请号：US10917063

申请日：2004-08-12

申请人： Mei Han ,  Yihong Gong ,  Hai Tao

发明人： Mei Han ,  Yihong Gong ,  Hai Tao

IPC分类号： G01S5/16 ,  G06K9/00 ,  G06T7/20 ,  G08C19/00 ,  H04N5/76 ,  H04N7/18

CPC分类号： G06T7/20 ,  G01S5/16 ,  G06K9/00295 ,  G06K9/00771 ,  G06K9/32 ,  G06T2207/30241 ,  H04N5/76 ,  H04N7/188

摘要： A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and/or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

摘要翻译： 视频监控系统使用基于规则的推理和多重假设评分来根据通过区域模式的移动来检测预定义的行为。 轨迹假设产卵允许轨迹分裂和/或合并，并包括局部修剪以管理假设增长。 假设基于许多标准进行评分，示例性地包括至少一个非空间参数。 在假设产卵过程期间计算的连接概率基于许多标准，示例性地包括对象大小。 对象检测和概率评分说明性地基于对象类。

公开(公告)号：US20050104962A1

公开(公告)日：2005-05-19

申请号：US10917985

申请日：2004-08-12

申请人： Mei Han ,  Yihong Gong ,  Hai Tao

发明人： Mei Han ,  Yihong Gong ,  Hai Tao

IPC分类号： G06K9/00 ,  H04N7/18 ,  H04N9/47

CPC分类号： G06K9/00771 ,  G06K9/00335 ,  G08B13/19608

摘要： A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and/or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

摘要翻译： 视频监控系统使用基于规则的推理和多重假设评分来根据通过区域模式的移动来检测预定义的行为。 轨迹假设产卵允许轨迹分裂和/或合并，并包括局部修剪以管理假设增长。 假设基于许多标准进行评分，示例性地包括至少一个非空间参数。 在假设产卵过程期间计算的连接概率基于许多标准，示例性地包括对象大小。 对象检测和概率评分说明性地基于对象类。

公开(公告)号：US20050104959A1

公开(公告)日：2005-05-19

申请号：US10916966

申请日：2004-08-12

申请人： Mei Han ,  Yihong Gong ,  Hai Tao

发明人： Mei Han ,  Yihong Gong ,  Hai Tao

IPC分类号： G08B13/196 ,  H04N9/47

CPC分类号： G08B13/19652 ,  G08B13/19604 ,  G08B13/19608 ,  G08B13/19615

摘要： A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and/or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

摘要翻译： 视频监控系统使用基于规则的推理和多重假设评分来根据通过区域模式的移动来检测预定义的行为。 轨迹假设产卵允许轨迹分裂和/或合并，并包括局部修剪以管理假设增长。 假设基于许多标准进行评分，示例性地包括至少一个非空间参数。 在假设产卵过程期间计算的连接概率基于许多标准，示例性地包括对象大小。 对象检测和概率评分说明性地基于对象类。

公开(公告)号：US06751776B1

公开(公告)日：2004-06-15

申请号：US09369421

申请日：1999-08-06

申请人： Yihong Gong

发明人： Yihong Gong

IPC分类号： G06F1500

CPC分类号： G06F17/30828 ,  G06F17/30616 ,  G06F17/30843

摘要： An automatic video content summarization system that is able to create personalized multimedia summary based on the user-specified theme. The invention employs both natural language processing and video analysis techniques to extract important keywords from the closed caption text as well as prominent visual features from the video footage. The invention uses a Bayesian statistical framework that naturally integrates the user theme, the heuristics and the theme-relevant video characteristics within a unified platform.

摘要翻译： 一种能够根据用户指定的主题创建个性化多媒体摘要的自动视频内容摘要系统。 本发明采用自然语言处理和视频分析技术，从隐藏的字幕文本中提取重要的关键词以及视频素材的突出的视觉特征。 本发明使用贝叶斯统计框架，其自然地将用户主题，启发式和与主题相关的视频特征集成在统一的平台内。