公开(公告)号：US07457432B2

公开(公告)日：2008-11-25

申请号：US11003004

申请日：2004-12-02

申请人： Haizhou Ai ,  Chang Huang ,  Bo Wu ,  Shihong Lao

发明人： Haizhou Ai ,  Chang Huang ,  Bo Wu ,  Shihong Lao

IPC分类号： G06K9/00 ,  A61B1/04

CPC分类号： G06K9/00228 ,  G06K9/6257

摘要： Unlike in the prior art in which the correspondence between a feature and a judgment value in an intended area is obtained by a single threshold value, the correspondence of the judgment value is obtained independently for each feature by use of a look-up table or the like. This makes it possible to achieve an accurate correspondence between the judgment value and the feature and thus to improve the high processing accuracy. Also, in the prior art, the judgment is repeated several times to secure the accuracy of the judgment and based on the total result thereof, the final judgment is made. Since the accuracy of each process is improved, however, the number of times the judgment is repeated is reduced for a higher processing speed.

摘要翻译： 与通过单个阈值获得特征与预期区域中的判断值之间的对应关系的现有技术不同，通过使用查找表或每个特征单独地获得每个特征的判断值的对应关系 喜欢。 这使得可以实现判断值与特征之间的精确对应，从而提高高处理精度。 此外，在现有技术中，重复多次判断以确保判断的准确性，并且基于其总结果，进行最终判断。 然而，由于每个处理的精度都得到改善，所以在更高的处理速度下减少了重复判断的次数。

公开(公告)号：US20050271245A1

公开(公告)日：2005-12-08

申请号：US11003004

申请日：2004-12-02

申请人： Haizhou Ai ,  Chang Huang ,  Bo Wu ,  Shihong Lao

发明人： Haizhou Ai ,  Chang Huang ,  Bo Wu ,  Shihong Lao

IPC分类号： G06T1/00 ,  G06K9/00 ,  G06K9/62 ,  G06T7/00

CPC分类号： G06K9/00228 ,  G06K9/6257

摘要： Unlike in the prior art in which the correspondence between a feature and a judgment value in an intended area is obtained by a single threshold value, the correspondence of the judgment value is obtained independently for each feature by use of a look-up table or the like. This makes it possible to achieve an accurate correspondence between the judgment value and the feature and thus to improve the high processing accuracy. Also, in the prior art, the judgment is repeated several times to secure the accuracy of the judgment and based on the total result thereof, the final judgment is made. Since the accuracy of each process is improved, however, the number of times the judgment is repeated is reduced for a higher processing speed.

摘要翻译： 与通过单个阈值获得特征与预期区域中的判断值之间的对应关系的现有技术不同，通过使用查找表或每个特征单独地获得每个特征的判断值的对应关系 喜欢。 这使得可以实现判断值与特征之间的精确对应，从而提高高处理精度。 此外，在现有技术中，重复多次判断以确保判断的准确性，并且基于其总结果，进行最终判断。 然而，由于每个处理的精度都得到改善，所以在更高的处理速度下减少了重复判断的次数。

公开(公告)号：US08457391B2

公开(公告)日：2013-06-04

申请号：US12602635

申请日：2008-05-30

申请人： Haizhou Ai ,  Chang Huang ,  Shihong Lao ,  Takayoshi Yamashita

发明人： Haizhou Ai ,  Chang Huang ,  Shihong Lao ,  Takayoshi Yamashita

IPC分类号： G06K9/62 ,  G06K9/54

CPC分类号： G06K9/6257 ,  G06K9/00248

摘要： The invention discloses a detecting device for specific subjects and a learning device and method thereof. The detecting device for specific subjects includes an input unit, one or more strong classifying units, a storage unit and a judging unit, wherein the input unit is used for inputting images to be detected; the strong classifying units are used for carrying out strong classification to the image, each strong classifying unit includes one or more weak classifying units, and the weak classifying unit carries out weak classification to the image with a weak classifying template; the storage unit stores the weak classifying template used by the weak classifying unit; and the judging unit judges whether or not the image contains specific subjects according to the classification result of the strong classifying unit. The detecting device for specific subjects also includes an incremental sample input unit and a learning unit, wherein the incremental sample input unit is used for inputting data for incremental learning, namely for inputting an incremental learning sample, which is data undetected and wrongly detected by the detecting device or other detecting devices for specific subjects; the learning unit is used for updating the weak classifying template stored in the storage unit according to the incremental learning sample inputted by the incremental sample input unit.

摘要翻译： 本发明公开了一种特定对象的检测装置及其学习装置及其方法。 用于特定对象的检测装置包括输入单元，一个或多个强分类单元，存储单元和判断单元，其中输入单元用于输入要检测的图像; 强分类单位用于对图像进行强分类，每个强分类单位包括一个或多个弱分类单位，弱分类单位用弱分类模板对图像进行弱分类; 存储单元存储由弱分类单元使用的弱分类模板; 并且判断单元根据强分类单元的分类结果判断图像是否包含特定对象。 用于特定主体的检测装置还包括增量采样输入单元和学习单元，其中增量采样输入单元用于输入用于增量学习的数据，即用于输入增量学习样本，该增量学习样本是未被检测到并被错误检测的数据 用于特定对象的检测装置或其他检测装置; 学习单元用于根据由增量抽样输入单元输入的递增学习样本来更新存储在存储单元中的弱分类模板。

公开(公告)号：US07876965B2

公开(公告)日：2011-01-25

申请号：US11543158

申请日：2006-10-05

申请人： Haizhou Ai ,  Chang Huang ,  Yuan Li ,  Shihong Lao

发明人： Haizhou Ai ,  Chang Huang ,  Yuan Li ,  Shihong Lao

IPC分类号： G06K9/70

CPC分类号： G06K9/6257 ,  G06K9/00208 ,  G06K9/00228 ,  G06K9/6282

摘要： An apparatus and a method for detecting from an image a particular subject corresponding to multiple views of the subject by dividing a particular subject space into a plurality of subject subspaces and further dividing a subject subspace into subject subspaces representing multiple views; configuring a tree-structured detector wherein the tree structure has a root node that covers all subject subspaces and has a plurality of branches, each branch corresponding to a child node that covers at least one subject subspace; training each node to determine which nodes in the adjacent lower layer the images of the subject in the corresponding nodes should be sent.

摘要翻译： 一种用于通过将特定对象空间划分成多个对象子空间并且将主题子空间划分成表示多个视图的对象子空间来检测与对象的多个视图相对应的特定对象的图像的装置和方法; 配置树结构检测器，其中所述树结构具有覆盖所有对象子空间并具有多个分支的根节点，每个分支对应于覆盖至少一个主题子空间的子节点; 训练每个节点以确定相邻下层中哪些节点应该发送相应节点中的对象的图像。

公开(公告)号：US08520955B2

公开(公告)日：2013-08-27

申请号：US12810449

申请日：2008-12-24

申请人： Haizhou Ai ,  Chang Huang ,  Shihong Lao ,  Takayoshi Yamashita

发明人： Haizhou Ai ,  Chang Huang ,  Shihong Lao ,  Takayoshi Yamashita

IPC分类号： G06K9/46

CPC分类号： G06K9/00221 ,  G06K9/00228 ,  G06K9/00248 ,  G06K9/4614

摘要： The present invention discloses an object detection apparatus and method. A feature extracting section of the present invention comprises: a feature point extracting section, for extracting a combination of predetermined feature point pairs from an image; a pixel value obtaining section, for obtaining a pixel value of each feature point in the combination of feature point pairs; a feature point comparing section, for comparing, in accordance with the pixel values obtained by the pixel value obtaining section, two feature points in each feature point pair to obtain a logical value; and an feature obtaining section, for determining the feature of the image in accordance with the logical value.

摘要翻译： 本发明公开了一种物体检测装置和方法。 本发明的特征提取部分包括：特征点提取部分，用于从图像中提取预定特征点对的组合; 像素值获取部，用于获得特征点对的组合中的每个特征点的像素值; 特征点比较部分，用于根据由像素值获取部分获得的像素值，比较每个特征点对中的两个特征点以获得逻辑值; 以及特征获取部分，用于根据该逻辑值确定图像的特征。

公开(公告)号：US20110007974A1

公开(公告)日：2011-01-13

申请号：US12810449

申请日：2008-12-24

申请人： Haizhou Ai ,  Chang Huang ,  Shihong Lao ,  Takayoshi Yamashita

发明人： Haizhou Ai ,  Chang Huang ,  Shihong Lao ,  Takayoshi Yamashita

IPC分类号： G06K9/46

CPC分类号： G06K9/00221 ,  G06K9/00228 ,  G06K9/00248 ,  G06K9/4614

摘要： The present invention discloses an object detection apparatus and method. A feature extracting section of the present invention comprises: a feature point extracting section, for extracting a combination of predetermined feature point pairs from an image; a pixel value obtaining section, for obtaining a pixel value of each feature point in the combination of feature point pairs; a feature point comparing section, for comparing, in accordance with the pixel values obtained by the pixel value obtaining section, two feature points in each feature point pair to obtain a logical value; and an feature obtaining section, for determining the feature of the image in accordance with the logical value.

摘要翻译： 本发明公开了一种物体检测装置和方法。 本发明的特征提取部分包括：特征点提取部分，用于从图像中提取预定特征点对的组合; 像素值获取部，用于获得特征点对的组合中的每个特征点的像素值; 特征点比较部分，用于根据由像素值获取部分获得的像素值，比较每个特征点对中的两个特征点以获得逻辑值; 以及特征获取部分，用于根据该逻辑值确定图像的特征。

公开(公告)号：US20070086660A1

公开(公告)日：2007-04-19

申请号：US11543158

申请日：2006-10-05

申请人： Haizhou Ai ,  Chang Huang ,  Yuan Li ,  Shihong Lao

发明人： Haizhou Ai ,  Chang Huang ,  Yuan Li ,  Shihong Lao

IPC分类号： G06K9/68

CPC分类号： G06K9/6257 ,  G06K9/00208 ,  G06K9/00228 ,  G06K9/6282

摘要： An apparatus and a method for detecting from an image a particular subject corresponding to multiple views of the subject by dividing a particular subject space into a plurality of subject subspaces and further dividing a subject subspace into subject subspaces representing multiple views; configuring a tree-structured detector wherein the tree structure has a root node that covers all subject subspaces and has a plurality of branches, each branch corresponding to a child node that covers at least one subject subspace; training each node to determine which nodes in the adjacent lower layer the images of the subject in the corresponding nodes should be sent.

摘要翻译： 一种用于通过将特定对象空间划分成多个对象子空间并且将主题子空间划分成表示多个视图的对象子空间来检测与对象的多个视图相对应的特定对象的图像的装置和方法; 配置树结构检测器，其中所述树结构具有覆盖所有对象子空间并具有多个分支的根节点，每个分支对应于覆盖至少一个主题子空间的子节点; 训练每个节点以确定相邻下层中哪些节点应该发送相应节点中的对象的图像。

公开(公告)号：US20100202681A1

公开(公告)日：2010-08-12

申请号：US12602635

申请日：2008-05-30

申请人： Haizhou Ai ,  Yuan Li ,  Shihong Lao ,  Takayoshi Yamashita

发明人： Haizhou Ai ,  Yuan Li ,  Shihong Lao ,  Takayoshi Yamashita

IPC分类号： G06K9/62

CPC分类号： G06K9/6257 ,  G06K9/00248

摘要： The invention discloses a detecting device for specific subjects and a learning device and method thereof. The detecting device for specific subjects includes an input unit, one or more strong classifying units, a storage unit and a judging unit, wherein the input unit is used for inputting images to be detected; the strong classifying units are used for carrying out strong classification to the image, each strong classifying unit includes one or more weak classifying units, and the weak classifying unit carries out weak classification to the image with a weak classifying template; the storage unit stores the weak classifying template used by the weak classifying unit; and the judging unit judges whether or not the image contains specific subjects according to the classification result of the strong classifying unit. The detecting device for specific subjects also includes an incremental sample input unit and a learning unit, wherein the incremental sample input unit is used for inputting data for incremental learning, namely for inputting an incremental learning sample, which is data undetected and wrongly detected by the detecting device or other detecting devices for specific subjects; the learning unit is used for updating the weak classifying template stored in the storage unit according to the incremental learning sample inputted by the incremental sample input unit.

摘要翻译： 本发明公开了一种特定对象的检测装置及其学习装置及其方法。 用于特定对象的检测装置包括输入单元，一个或多个强分类单元，存储单元和判断单元，其中输入单元用于输入要检测的图像; 强分类单位用于对图像进行强分类，每个强分类单位包括一个或多个弱分类单位，弱分类单位用弱分类模板对图像进行弱分类; 存储单元存储由弱分类单元使用的弱分类模板; 并且判断单元根据强分类单元的分类结果判断图像是否包含特定对象。 用于特定主体的检测装置还包括增量采样输入单元和学习单元，其中增量采样输入单元用于输入用于增量学习的数据，即用于输入增量学习样本，该增量学习样本是未被检测到并被错误检测的数据 用于特定对象的检测装置或其他检测装置; 学习单元用于根据由增量抽样输入单元输入的递增学习样本来更新存储在存储单元中的弱分类模板。

公开(公告)号：US20100195870A1

公开(公告)日：2010-08-05

申请号：US12664588

申请日：2008-06-13

申请人： Haizhou Ai ,  Yuan Li ,  Shihong Lao ,  Takayoshi Yamashita

发明人： Haizhou Ai ,  Yuan Li ,  Shihong Lao ,  Takayoshi Yamashita

IPC分类号： G06K9/62

CPC分类号： G06T7/277 ,  G06T2207/10016 ,  G06T2207/30201

摘要： The present invention relates to a tracking method and a tracking device adopting multiple observation models with different life spans. The tracking method is suitable for tracking an object in a low frame rate video or with abrupt motion, and uses three observation models with different life spans to track and detect a specific subject in frame images of a video sequence. An observation model I performs online learning with one frame image prior to the current image, an observation model II performs online learning with five frames prior to the current image, and an observation model III is offline trained. The three observation models are combined by a cascade particle filter so that the specific subject in the low frame rate video or the object with abrupt motion can be tracked quickly and accurately.

摘要翻译： 本发明涉及采用具有不同寿命的多个观察模型的跟踪方法和跟踪装置。 跟踪方法适用于跟踪低帧率视频或突然运动的对象，并且使用具有不同寿命的三个观察模型来跟踪和检测视频序列的帧图像中的特定主体。 观察模型I在当前图像之前用一帧图像执行在线学习，观察模型II在当前图像之前用五帧执行在线学习，并且观察模型III被离线训练。 三个观察模型通过级联粒子滤波器组合，使得低帧率视频中的特定对象或具有突然运动的对象能够被快速准确地跟踪。

公开(公告)号：US20080063263A1

公开(公告)日：2008-03-13

申请号：US11517485

申请日：2006-09-08

申请人： Li Zhang ,  Haizhou Ai ,  Shuichiro Tsukiji ,  Shihong Lao

发明人： Li Zhang ,  Haizhou Ai ,  Shuichiro Tsukiji ,  Shihong Lao

IPC分类号： G06K9/62 ,  G06K9/46 ,  G06K9/66

CPC分类号： G06K9/00234 ,  G06K9/00248 ,  G06K9/6209

摘要： Method and apparatus for face alignment by building a hierarchical classifier network. The hierarchical classifier network connects the tasks of face detection and face alignment into a smooth coarse-to-fine procedure. Texture classifiers are trained to recognize feature texture at different scales for different resolution layers. A multi-layer structure is employed to organize the texture classifiers, which begins with one classifier at the first layer and gradually refines the localization of feature points using additional texture classifiers in subsequent layers.

摘要翻译： 通过构建层次分类器网络进行面对齐的方法和装置。 分层分类器网络将面部检测和面部对准的任务连接成一个平滑的粗略到精细的过程。 训练纹理分类器以识别不同分辨率层的不同尺度的特征纹理。 使用多层结构来组织纹理分类器，其从第一层的一个分类器开始，并且使用后续层中的附加纹理分类器逐渐地精化特征点的定位。