公开(公告)号：US07466755B2

公开(公告)日：2008-12-16

申请号：US10410274

申请日：2003-04-10

申请人： Tsuhan Chen ,  Trista Pei-Chun Chen ,  Yuh-Feng Hsu

发明人： Tsuhan Chen ,  Trista Pei-Chun Chen ,  Yuh-Feng Hsu

IPC分类号： H04N7/12 ,  H04N11/02

CPC分类号： H04N19/895

摘要： A method of error concealment that includes providing a first set of image data at a first time instant, providing a second set of image data at a second time instant, representing the first set of image data by a model of updating mixture of principal components (UMPC) to form a UMPC model having a plurality of components, receiving the second set of image data, wherein the second set of image data contains error-free data, and updating the UMPC model using the error-free data at the second time instant.

摘要翻译： 一种错误隐藏的方法，包括在第一时刻提供第一组图像数据，在第二时刻提供第二组图像数据，通过更新主成分混合的模型来表示第一组图像数据（ UMPC）以形成具有多个组件的UMPC模型，接收第二组图像数据，其中第二组图像数据包含无错误数据，并且在第二时间点使用无差错数据更新UMPC模型 。

公开(公告)号：US06944319B1

公开(公告)日：2005-09-13

申请号：US09536820

申请日：2000-03-27

申请人： Fu Jie Huang ,  Hong-Jiang Zhang ,  Tsuhan Chen

发明人： Fu Jie Huang ,  Hong-Jiang Zhang ,  Tsuhan Chen

IPC分类号： G06K9/00 ,  G06K9/62 ,  G06K9/68

CPC分类号： G06K9/6282 ,  G06K9/00288

摘要： A face recognition system and process for identifying a person depicted in an input image and their face pose. This system and process entails locating and extracting face regions belonging to known people from a set of model images, and determining the face pose for each of the face regions extracted. All the extracted face regions are preprocessed by normalizing, cropping, categorizing and finally abstracting them. More specifically, the images are normalized and cropped to show only a person's face, categorized according to the face pose of the depicted person's face by assigning them to one of a series of face pose ranges, and abstracted preferably via an eigenface approach. The preprocessed face images are preferably used to train a neural network ensemble having a first stage made up of a bank of face recognition neural networks each of which is dedicated to a particular pose range, and a second stage constituting a single fusing neural network that is used to combine the outputs from each of the first stage neural networks. Once trained, the input of a face region which has been extracted from an input image and preprocessed (i.e., normalized, cropped and abstracted) will cause just one of the output units of the fusing portion of the neural network ensemble to become active. The active output unit indicates either the identify of the person whose face was extracted from the input image and the associated face pose, or that the identity of the person is unknown to the system.

摘要翻译： 用于识别输入图像中描绘的人物及其脸部姿势的面部识别系统和过程。 该系统和处理需要从一组模型图像中定位和提取属于已知人的面部区域，并且确定提取的每个面部区域的面部姿势。 所有提取的面部区域通过归一化，裁剪，分类和最终抽象来进行预处理。 更具体地，图像被归一化并且被裁剪以仅显示人脸，根据所描绘的人脸的脸部姿态将其分配给一系列面部姿态范围中的一个，并优选地通过本征面方法进行抽象。 预处理的脸部图像优选地用于训练具有由一组专用于特定姿势范围的面部识别神经网络组成的第一阶段的神经网络集合，以及构成单个融合神经网络的第二阶段， 用于组合来自每个第一级神经网络的输出。 一旦被训练，已经从输入图像提取并且被预处理（即，归一化，裁剪和抽取）的面部区域的输入将仅导致神经网络集合的融合部分的输出单元中的一个变得活动。 活动输出单元指示从输入图像中提取脸部的人的身份以及相关联的面部姿势，或者该人的身份对于系统是未知的。

公开(公告)号：US20050147291A1

公开(公告)日：2005-07-07

申请号：US10982743

申请日：2004-11-05

申请人： Fu Huang ,  Hong-Jiang Zhang ,  Tsuhan Chen

发明人： Fu Huang ,  Hong-Jiang Zhang ,  Tsuhan Chen

IPC分类号： G06K9/00 ,  G06K9/62 ,  G06K9/68

CPC分类号： G06K9/6282 ,  G06K9/00288

摘要： A face recognition system and process for identifying a person depicted in an input image and their face pose. This system and process entails locating and extracting face regions belonging to known people from a set of model images, and determining the face pose for each of the face regions extracted. All the extracted face regions are preprocessed by normalizing, cropping, categorizing and finally abstracting them. More specifically, the images are normalized and cropped to show only a persons face, categorized according to the face pose of the depicted person's face by assigning them to one of a series of face pose ranges, and abstracted preferably via an eigenface approach. The preprocessed face images are preferably used to train a neural network ensemble having a first stage made up of a bank of face recognition neural networks each of which is dedicated to a particular pose range, and a second stage constituting a single fusing neural network that is used to combine the outputs from each of the first stage neural networks. Once trained, the input of a face region which has been extracted from an input image and preprocessed (i.e., normalized, cropped and abstracted) will cause just one of the output units of the fusing portion of the neural network ensemble to become active. The active output unit indicates either the identify of the person whose face was extracted from the input image and the associated face pose, or that the identity of the person is unknown to the system.

摘要翻译： 用于识别输入图像中描绘的人物及其脸部姿势的面部识别系统和过程。 该系统和处理需要从一组模型图像中定位和提取属于已知人的面部区域，并且确定提取的每个面部区域的面部姿势。 所有提取的面部区域通过归一化，裁剪，分类和最终抽象来进行预处理。 更具体地，将图像归一化并裁剪以仅显示人脸，根据所描绘的人脸的脸部姿态将其分配给一系列面部姿势范围中的一个，并且优选地通过特征面方法进行抽象。 预处理的脸部图像优选地用于训练具有由一组专用于特定姿势范围的面部识别神经网络组成的第一阶段的神经网络集合，以及构成单个融合神经网络的第二阶段， 用于组合来自每个第一级神经网络的输出。 一旦被训练，已经从输入图像提取并且被预处理（即，归一化，裁剪和抽取）的面部区域的输入将仅导致神经网络集合的融合部分的输出单元中的一个变得活动。 活动输出单元指示从输入图像中提取脸部的人的身份以及相关联的面部姿势，或者该人的身份对于系统是未知的。

公开(公告)号：US06330023B1

公开(公告)日：2001-12-11

申请号：US08210529

申请日：1994-03-18

申请人： Tsuhan Chen

发明人： Tsuhan Chen

IPC分类号： H04N713

CPC分类号： H04N21/4341 ,  G10L2021/105 ,  H04N19/132 ,  H04N19/20 ,  H04N19/587 ,  H04N21/4394 ,  H04N21/44008 ,  H04N21/440281

摘要： A method of increasing the frame rate of an image of a speaking person comprises monitoring an audio signal indicative of utterances by the speaking person and the associated video signal. The audio signal corresponds to one or more fields or frames to be reconstructed, and individual portions of the audio signal are associated with facial feature information. The facial information includes mouth formation and position information derived from phonemes or other speech-based criteria from which the position of a speaker's mouth may be reliably predicted. A field or frame of the image is reconstructed using image features extracted from the existing frame and by utilizing the facial feature information associated with a detected phoneme.

摘要翻译： 提高说话人的图像的帧速率的方法包括监视由说话人和相关联的视频信号指示话语的音频信号。 音频信号对应于要重建的一个或多个字段或帧，并且音频信号的各个部分与面部特征信息相关联。 面部信息包括口音形成和从音素或其他基于语音的标准导出的位置信息，从该位置信息可以可靠地预测说话者的嘴的位置。 使用从现有帧提取的图像特征并利用与检测到的音素相关联的面部特征信息来重建图像的场或帧。

公开(公告)号：US6141442A

公开(公告)日：2000-10-31

申请号：US358594

申请日：1999-07-21

申请人： Tsuhan Chen

发明人： Tsuhan Chen

IPC分类号： H04N7/26 ,  H04N9/75 ,  G06K9/00

CPC分类号： H04N9/75 ,  G06T9/001 ,  G06T9/20 ,  H04N19/186

摘要： A method and apparatus for generating region frames from video frames are disclosed which employs an industry standard encoder to lessen the negative impact on the quality of the transmitted video sequence while consuming fewer bits. The invention utilizes image segmentation and color replacement techniques to create the region frames. Each region frame includes a subject region, zero or more previously segmented regions and zero or more non-subject regions. The subject region is defined by the pixels of the original video frame. The previously segmented regions and non-subject regions are assigned replacement pixels P.sub.n,y and C.sub.n, respectively. The replacement pixel C.sub.n is chosen to indicate a color that is not likely to be confused with any color in the subject region R.sub.n. The replacement pixels P.sub.n,y are chosen such that the compression ratio of the region frame data is maximized. Using the region frames, content based scalability can be provided without the need for special encoders and/or channels having a wider bandwidth. The decoder may comprise color or chroma keying apparatus or circuitry keying on the replacement color C.sub.n. Instead of keying on a single value, two thresholds may be assigned to define a boundary condition or a subject semi-transparent region. The decoder is forwarded data of the two thresholds and a flag is sent to indicate the dial boundary or semi-transparent region coding. A blending process blends the foreground and background of the semi-transparent object.

摘要翻译： 公开了一种用于从视频帧产生区域帧的方法和装置，其采用工业标准编码器来减少对所发送的视频序列的质量的负面影响，同时消耗较少的比特。 本发明利用图像分割和颜色替换技术来创建区域框架。 每个区域帧包括主题区域，零个或多个先前分割的区域和零个或多个非对象区域。 主题区域由原始视频帧的像素定义。 先前分割的区域和非对象区域分别被分配替换像素Pn，y和Cn。 选择替换像素Cn以指示不可能与对象区域Rn中的任何颜色混淆的颜色。 选择替换像素Pn，y使得区域帧数据的压缩比最大化。 使用区域帧，可以提供基于内容的可扩展性，而不需要具有更宽带宽的特殊编码器和/或信道。 解码器可以包括颜色或色度键控设备或键入替换颜色Cn的电路。 可以分配两个阈值来定义边界条件或对象半透明区域，而不是键入单个值。 解码器是两个阈值的转发数据，并且发送标志以指示拨号边界或半透明区域编码。 混合过程混合了半透明对象的前景和背景。

公开(公告)号：US5786855A

公开(公告)日：1998-07-28

申请号：US548818

申请日：1995-10-26

申请人： Tsuhan Chen ,  Barin Geoffry Haskell

发明人： Tsuhan Chen ,  Barin Geoffry Haskell

IPC分类号： H04N7/26 ,  H04N7/30 ,  H04N7/46 ,  H04N7/12

CPC分类号： H04N19/649 ,  H04N19/132 ,  H04N19/587 ,  H04N19/85 ,  H04N19/39

摘要： A method and apparatus for generating region frames from video frames are disclosed which employs an industry standard encoder to lessen the negative impact on the quality of the transmitted video sequence while consuming fewer bits. The method and apparatus utilizes image segmentation and color replacement techniques to create the region frames. Each region frame includes a subject region, zero or more previously segmented regions and zero or more non-subject regions. The subject region is defined by the pixels of the original video frame. The previously segmented regions and non-subject regions are assigned replacement pixels P.sub.n,y and C.sub.n, respectively. The replacement pixel C.sub.n is chosen to indicate a color that is not likely to be confused with any color in the subject region R.sub.n. The replacement pixels P.sub.n,y are chosen such that the compression ratio of the region frame data is maximized. Using the region frames, content based scalability can be provided without the need for special encoders and/or channels having a wider bandwidth.

摘要翻译： 公开了一种用于从视频帧产生区域帧的方法和装置，其采用工业标准编码器来减少对所发送的视频序列的质量的负面影响，同时消耗较少的比特。 该方法和装置利用图像分割和颜色替换技术来创建区域帧。 每个区域帧包括主题区域，零个或多个先前分割的区域和零个或多个非对象区域。 主题区域由原始视频帧的像素定义。 先前分割的区域和非对象区域分别被分配替换像素Pn，y和Cn。 选择替换像素Cn以指示不可能与对象区域Rn中的任何颜色混淆的颜色。 选择替换像素Pn，y使得区域帧数据的压缩比最大化。 使用区域帧，可以提供基于内容的可扩展性，而不需要具有更宽带宽的特殊编码器和/或信道。

公开(公告)号：US08487991B2

公开(公告)日：2013-07-16

申请号：US12581659

申请日：2009-10-19

申请人： Wende Zhang ,  Qi Wu ,  Tsuhan Chen

发明人： Wende Zhang ,  Qi Wu ,  Tsuhan Chen

IPC分类号： H04N7/00 ,  H04N7/18

CPC分类号： G06K9/00798 ,  B60W2420/42

摘要： A method for estimating a vanishing point in a roadway using a current image generated by a camera on a vehicle includes defining an exemplary vanishing point for each of a plurality of sample images, identifying features within each of the plurality of sample images, monitoring the current image generated by the camera, identifying features within the current image, matching the current image to at least one of the sample images based upon the identified features within the current image and the identified features within the plurality of sample images, determining a vanishing point based upon the matching and the exemplary vanishing points for each of the matched sample images, and utilizing the vanishing point to navigate the vehicle.

摘要翻译： 使用车辆上的照相机产生的当前图像来估计道路中的消失点的方法包括为多个样本图像中的每一个定义示例性消失点，识别多个样本图像中的每一个内的特征，监视当前 由相机生成的图像，识别当前图像内的特征，基于当前图像内的所识别的特征以及多个样本图像中识别的特征，将当前图像与至少一个样本图像匹配，基于消失点 对于每个匹配的样本图像的匹配和示例性消失点，并且利用消失点来导航车辆。

公开(公告)号：US07127087B2

公开(公告)日：2006-10-24

申请号：US10983194

申请日：2004-11-05

申请人： Fu Jie Huang ,  Hong-Jiang Zhang ,  Tsuhan Chen

发明人： Fu Jie Huang ,  Hong-Jiang Zhang ,  Tsuhan Chen

IPC分类号： G06K9/00 ,  G06K9/62

CPC分类号： G06K9/00228 ,  G06K9/00288 ,  G06K9/6292

摘要： A face recognition system and process for identifying a person depicted in an input image and their face pose. This system and process entails locating and extracting face regions belonging to known people from a set of model images, and determining the face pose for each of the face regions extracted. All the extracted face regions are preprocessed by normalizing, cropping, categorizing and finally abstracting them. More specifically, the images are normalized and cropped to show only a persons face, categorized according to the face pose of the depicted person's face by assigning them to one of a series of face pose ranges, and abstracted preferably via an eigenface approach. The preprocessed face images are preferably used to train a neural network ensemble having a first stage made up of a bank of face recognition neural networks each of which is dedicated to a particular pose range, and a second stage constituting a single fusing neural network that is used to combine the outputs from each of the first stage neural networks. Once trained, the input of a face region which has been extracted from an input image and preprocessed (i.e., normalized, cropped and abstracted) will cause just one of the output units of the fusing portion of the neural network ensemble to become active. The active output unit indicates either the identify of the person whose face was extracted from the input image and the associated face pose, or that the identity of the person is unknown to the system.

公开(公告)号：US6058187A

公开(公告)日：2000-05-02

申请号：US842784

申请日：1997-04-17

申请人： Tsuhan Chen

发明人： Tsuhan Chen

IPC分类号： H04L9/08 ,  H04L29/06 ,  H04M7/12 ,  H04L9/00 ,  H04L9/30 ,  H04L9/32

CPC分类号： H04L63/0442 ,  H04L29/06 ,  H04L63/061 ,  H04L9/14 ,  H04M7/12

摘要： Secure data transmission apparatus comprises a data translator for translating an input string of signals, each signal having incomplete information for identifying an alphanumeric character, into a first encryption key. A data encrypter receives a first encryption key, a choice of encryption algorithm and a message and outputs an encrypted message according to the selected algorithm. The apparatus may be applied whenever the user is confronted with telecommunications apparatus that provides a limited input capability and no means for encrypting a message for transmission to an end user. In this manner, for example, a user may authenticate their name for display on caller identification plus name apparatus and the called party can be assured, before answering their line, that the call is from the party having the displayed name.

摘要翻译： 安全数据传输装置包括数据转换器，用于将每个信号的输入串转换成第一加密密钥，每个信号具有用于识别字母数字字符的不完整信息。 数据加密器接收第一加密密钥，选择加密算法和消息，并根据所选算法输出加密消息。 只要用户面对提供有限输入能力的电信设备，并且没有用于加密消息以传送给最终用户的装置，则可以应用该设备。 以这种方式，例如，用户可以认证其姓名以在呼叫者身份加上名称设备上显示，并且被叫方可以在应答他们的线路之前确保该呼叫是来自具有所显示姓名的一方。

公开(公告)号：US5974172A

公开(公告)日：1999-10-26

申请号：US801716

申请日：1997-02-14

申请人： Tsuhan Chen

发明人： Tsuhan Chen

IPC分类号： H04N7/26 ,  H04N9/75 ,  G06K9/00

CPC分类号： H04N9/75 ,  G06T9/001 ,  G06T9/20 ,  H04N19/186

摘要： A method and apparatus for generating region frames from video frames are disclosed which employs an industry standard encoder to lessen the negative impact on the quality of the transmitted video sequence while consuming fewer bits. The invention utilizes image segmentation and color replacement techniques to create the region frames. Each region frame includes a subject region, zero or more previously segmented regions and zero or more non-subject regions. The subject region is defined by the pixels of the original video frame. The previously segmented regions and non-subject regions are assigned replacement pixels P.sub.n,y and C.sub.n, respectively. The replacement pixel C.sub.n is chosen to indicate a color that is not likely to be confused with any color in the subject region R.sub.n. The replacement pixels P.sub.n,y are chosen such that the compression ratio of the region frame data is maximized. Using the region frames, content based scalability can be provided without the need for special encoders and/or channels having a wider bandwidth. The decoder may comprise color or chroma keying apparatus or circuitry keying on the replacement color C.sub.n. Instead of keying on a single value, two thresholds may be assigned to define a boundary condition or a subject semi-transparent region. The decoder is forwarded data of the two thresholds and a flag is sent to indicate the special boundary or semi-transparent region coding. A blending process blends the foreground and background of the semi-transparent object.

摘要翻译： 公开了一种用于从视频帧产生区域帧的方法和装置，其采用工业标准编码器来减少对所发送的视频序列的质量的负面影响，同时消耗较少的比特。 本发明利用图像分割和颜色替换技术来创建区域框架。 每个区域帧包括主题区域，零个或多个先前分割的区域和零个或多个非对象区域。 主题区域由原始视频帧的像素定义。 先前分割的区域和非对象区域分别被分配替换像素Pn，y和Cn。 选择替换像素Cn以指示不可能与对象区域Rn中的任何颜色混淆的颜色。 选择替换像素Pn，y使得区域帧数据的压缩比最大化。 使用区域帧，可以提供基于内容的可扩展性，而不需要具有更宽带宽的特殊编码器和/或信道。 解码器可以包括颜色或色度键控设备或键入替换颜色Cn的电路。 可以分配两个阈值来定义边界条件或对象半透明区域，而不是键入单个值。 解码器是两个阈值的转发数据，并且发送标志以指示特殊边界或半透明区域编码。 混合过程混合了半透明对象的前景和背景。