公开(公告)号：US08571322B2

公开(公告)日：2013-10-29

申请号：US12546376

申请日：2009-08-24

申请人： Kozo Baba ,  Akiyoshi Tafuku ,  Hisao Ito ,  Jou Tanji

发明人： Kozo Baba ,  Akiyoshi Tafuku ,  Hisao Ito ,  Jou Tanji

IPC分类号： G06K9/00

CPC分类号： G06K9/00248

摘要： An image processing method for calculating feature amounts of a facial part in a face on the basis of a plurality of image data obtained in chronological order, comparing the calculated feature amounts with a threshold value, and recognizing the facial part, includes: calculating face orientations on the basis of image data; storing, in a storage, the feature amounts calculated from the image data, the feature amounts being associated with the face orientations; and recognizing a facial part in different image data of the face other than the plurality of the image data on the basis of a feature amount calculated from the different image data, the feature amounts stored in the storage associated with the face orientation in the different image data, and a threshold.

摘要翻译： 一种图像处理方法，用于根据以时间顺序获得的多个图像数据计算面部中的面部部分的特征量，将所计算的特征量与阈值进行比较，并识别面部部分，包括：计算面部取向 在图像数据的基础上; 在存储器中存储从所述图像数据计算的特征量，所述特征量与所述面部取向相关联; 以及基于从不同图像数据计算出的特征量，识别除了多个图像数据之外的面部的不同图像数据中的面部部分，存储在与不同图像中的面部取向相关联的存储器中的特征量 数据和阈值。

公开(公告)号：US08045765B2

公开(公告)日：2011-10-25

申请号：US11889999

申请日：2007-08-17

申请人： Kozo Baba ,  Hisao Ito ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Taku Katagiri

发明人： Kozo Baba ,  Hisao Ito ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Taku Katagiri

IPC分类号： G06K9/00

CPC分类号： G06K9/00248 ,  G06T7/74 ,  G06T2207/30201

摘要： For example, a plurality of elliptical candidate areas that become candidate areas of detecting objects are set in an image, characteristic values such as average values of luminance of the pixels inside and outside the set candidate areas are compared, and a candidate area whose difference of the characteristic values is determined to be the maximum is detected as the area of the detecting object. In addition, central moment of the detecting object is calculated, and the inclination of the detecting object in the image is detected based on the calculated central moment. Further, two specific points such as eyes of the driver included in the detecting object are detected, and the inclination of the detecting object in the image determined by the detected two points is detected.

摘要翻译： 例如，将成为检测对象的候补区域的多个椭圆候选区域设定为图像，比较设定候选区域内外的像素的亮度的平均值等特征值， 将特征值确定为最大值被检测为检测对象的区域。 此外，计算检测对象的中心矩，并且基于计算出的中心时刻来检测图像中的检测对象的倾斜度。 此外，检测包括在检测对象中的驱动器的眼睛的两个特定点，并且检测由检测到的两个点确定的图像中的检测对象的倾斜度。

公开(公告)号：US07991202B2

公开(公告)日：2011-08-02

申请号：US11889994

申请日：2007-08-17

申请人： Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Taku Katagiri

发明人： Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Taku Katagiri

IPC分类号： G06K9/00

CPC分类号： G06K9/00234 ,  G06T7/73 ,  G06T2207/30201

摘要： There are provided an image processing method, and an image processing system, an image processing device, and a computer program product capable of detecting a position of a fringe portion of a detecting object such as an outline of a face of a driver with high precision, even when a variation of illuminance occurs locally in a system using, for example, an on-vehicle camera mounted on a vehicle for imaging the face of the driver. Luminance of pixels arranged in a first direction such as a horizontal direction of an image is subjected to quadric differential, and data based on a quadratic differentiation result is accumulated in a second direction such as a vertical direction, and based on an accumulated result, a position in the first direction of the detecting object in the image is detected.

摘要翻译： 提供了一种图像处理方法，以及能够以高精度检测驾驶员的面部轮廓的检测对象的边缘部分的位置的图像处理系统，图像处理装置和计算机程序产品 即使在使用例如安装在车辆上的用于对驾驶员的脸进行成像的车载摄像机的系统中本发明的照度变化也是如此。 排列在图像的水平方向等第一方向上的像素的亮度经受二次微分，并且基于二次微分结果的数据在诸如垂直方向的第二方向上累积，并且基于累积结果， 检测图像中检测对象的第一方向的位置。

公开(公告)号：US20090310829A1

公开(公告)日：2009-12-17

申请号：US12546376

申请日：2009-08-24

申请人： Kozo Baba ,  Akiyoshi Tafuku ,  Hisao Ito ,  Jou Tanji

发明人： Kozo Baba ,  Akiyoshi Tafuku ,  Hisao Ito ,  Jou Tanji

IPC分类号： G06K9/00

CPC分类号： G06K9/00248

摘要： An image processing method for calculating feature amounts of a facial part in a face on the basis of a plurality of image data obtained in chronological order, comparing the calculated feature amounts with a threshold value, and recognizing the facial part, includes: calculating face orientations on the basis of image data; storing, in a storage, the feature amounts calculated from the image data, the feature amounts being associated with the face orientations; and recognizing a facial part in different image data of the face other than the plurality of the image data on the basis of a feature amount calculated from the different image data, the feature amounts stored in the storage associated with the face orientation in the different image data, and a threshold.

摘要翻译： 一种图像处理方法，用于根据以时间顺序获得的多个图像数据计算面部中的面部部分的特征量，将所计算的特征量与阈值进行比较，并识别面部部分，包括：计算面部取向 在图像数据的基础上; 在存储器中存储从所述图像数据计算的特征量，所述特征量与所述面部取向相关联; 以及基于从不同图像数据计算出的特征量，识别除了多个图像数据之外的面部的不同图像数据中的面部部分，存储在与不同图像中的面部取向相关联的存储器中的特征量 数据和阈值。

公开(公告)号：US20070291999A1

公开(公告)日：2007-12-20

申请号：US11844097

申请日：2007-08-23

申请人： Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku ,  Taku Saito ,  Masatoshi Tohno ,  Taku Katagiri

发明人： Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku ,  Taku Saito ,  Masatoshi Tohno ,  Taku Katagiri

IPC分类号： G06K9/00

CPC分类号： G06K9/00832 ,  G06K9/00248 ,  G06K9/4647 ,  G06T7/73 ,  G06T2207/30201

摘要： Provided are an image processing method, an image processing system, an image processing device and a computer program for detecting a detection object such as nares of the driver with a high degree of accuracy in, for example, a system using an in-vehicle camera which is mounted on a vehicle and takes an image of the face of the driver. A detection object is diversified by a variety of detection methods such as a method for detecting a plurality of locations in the vertical direction as candidates during image pickup of an image, detecting a range to be a candidate of a detection object on the basis of the brightness of a pixel for each of rows of pixels lined up in the horizontal direction corresponding to each detected location and specifying a detection object from candidates of a detection object on the basis of the length of the detected range.

摘要翻译： 提供了一种图像处理方法，图像处理系统，图像处理装置和计算机程序，用于在例如使用车载摄像机的系统中以高精度检测驾驶员的鼻孔等检测对象 其安装在车辆上并拍摄驾驶员的脸部的图像。 检测对象通过各种检测方法多样化，例如用于在垂直方向上检测多个位置的方法，作为图像的图像拾取期间的候选，基于检测对象的检测对象的候选值检测范围 根据检测范围的长度，在与检测到的各个位置对应的水平方向排列的每行像素的像素的亮度，并根据检测对象的长度来指定检测对象。

公开(公告)号：US07218213B2

公开(公告)日：2007-05-15

申请号：US11135431

申请日：2005-05-24

申请人： Taku Katagiri ,  Tomonobu Takashima ,  Masatoshi Tohno ,  Daisuke Ueno ,  Tadao Omi ,  Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku

发明人： Taku Katagiri ,  Tomonobu Takashima ,  Masatoshi Tohno ,  Daisuke Ueno ,  Tadao Omi ,  Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku

IPC分类号： B60Q1/54

CPC分类号： G08B21/06 ,  B60K28/04 ,  B60K28/06 ,  B60W40/09 ,  B60W2520/10 ,  B60W2520/105 ,  B60W2520/125 ,  B60W2530/18

摘要： An alarm system includes a miniature camera (10), a loudspeaker (20), a vehicle speed sensor (30), a longitudinal G sensor (40), a lateral G sensor (50) and an on-vehicle unit (60). A CPU (60a) of the on-vehicle unit (60) executes respective programs (61)–(64) in a ROM (60h), thereby outputting, if a traveling distance during a period for which a driver D continues to take an abnormal behavior exceeds a predetermined value when a traveling state of an automobile is kept stable, an alarm corresponding to a magnitude of the excess over the predetermined value from the loudspeaker (20). The alarm system is thus configured and is therefore capable of exactly outputting the alarm when a danger actually rises.

摘要翻译： 报警系统包括微型照相机（10），扬声器（20），车速传感器（30），纵向G传感器（40），横向G传感器（50）和车载单元（60）。 车载单元（60）的CPU（60a）在ROM（60h）中执行各自的程序（61） - （64），从而如果在驾驶员D继续期间的行驶距离 当汽车的行驶状态保持稳定时，超出预定值的异常行为，与来自扬声器（20）的超过预定值的超出量的警报相对应。 因此，报警系统被配置，并且因此能够在危险实际上升时精确地输出警报。

公开(公告)号：US08077916B2

公开(公告)日：2011-12-13

申请号：US11892140

申请日：2007-08-20

申请人： Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Tadao Omi ,  Taku Katagiri

发明人： Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Tadao Omi ,  Taku Katagiri

IPC分类号： G06K9/00

CPC分类号： G06K9/2027 ,  G06K9/00228

摘要： Using an on-vehicle camera mounted on a vehicle to view the face of a driver as an imaging object, an image processing system judges the presence/absence of the imaging object in the image. An average luminance of an image when light of a particular wavelength is irradiated in an imaging direction is compared to an average luminance of an image imaged when no light is irradiated. When the difference between the average luminance values is equal to a reference value or less, it is judged that the imaging object is absent in the image by imaging.

摘要翻译： 使用安装在车辆上的车载摄像机将驾驶员的脸部视为成像对象，图像处理系统判断图像中成像对象的存在/不存在。 将特定波长的光在成像方向照射时的平均亮度与未照射光时成像的图像的平均亮度进行比较。 当平均亮度值之间的差值等于或小于参考值时，通过成像判断图像中不存在成像对象。

公开(公告)号：US20070292001A1

公开(公告)日：2007-12-20

申请号：US11889999

申请日：2007-08-17

申请人： Kozo Baba ,  Hisao Ito ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Taku Katagiri

发明人： Kozo Baba ,  Hisao Ito ,  Akiyoshi Tafuku ,  Masatoshi Tohno ,  Taku Katagiri

IPC分类号： G06K9/00

CPC分类号： G06K9/00248 ,  G06T7/74 ,  G06T2207/30201

摘要： For example, a plurality of elliptical candidate areas that become candidate areas of detecting objects are set in an image, characteristic values such as average values of luminance of the pixels inside and outside the set candidate areas are compared, and a candidate area whose difference of the characteristic values is determined to be the maximum is detected as the area of the detecting object. In addition, central moment of the detecting object is calculated, and the inclination of the detecting object in the image is detected based on the calculated central moment. Further, two specific points such as eyes of the driver included in the detecting object are detected, and the inclination of the detecting object in the image determined by the detected two points is detected.

摘要翻译： 例如，将成为检测对象的候补区域的多个椭圆候选区域设定为图像，比较设定候选区域内外的像素的亮度的平均值等特征值， 将特征值确定为最大值被检测为检测对象的区域。 此外，计算检测对象的中心矩，并且基于计算出的中心时刻来检测图像中的检测对象的倾斜度。 此外，检测包括在检测对象中的驱动器的眼睛的两个特定点，并且检测由检测到的两个点确定的图像中的检测对象的倾斜度。

公开(公告)号：US07209574B2

公开(公告)日：2007-04-24

申请号：US10763209

申请日：2004-01-26

申请人： Akiyoshi Tafuku ,  Kozo Baba ,  Hisao Ito ,  Tomonobu Takashima ,  Tadao Omi

发明人： Akiyoshi Tafuku ,  Kozo Baba ,  Hisao Ito ,  Tomonobu Takashima ,  Tadao Omi

IPC分类号： G06K9/00

CPC分类号： G06K9/00604 ,  G06T7/215 ,  G06T7/248 ,  G06T2207/10016 ,  G06T2207/30201 ,  G06T2207/30248 ,  G06T2207/30268

摘要： Obtaining an image frame picked up by a camera 2 through a cable C, a CPU 10 of a computer 1 identifies an eye area in the image frame and stores a template indicating the identified area as a tracking template. Moreover, the CPU 10 identifies an eye area in an image frame to be obtained subsequently on the basis of the tracking template, and calculates a correlation value between the template of the identified area and the tracking template. Furthermore, when the calculated correlation value is larger than a threshold stored in an HD 11 in advance, the CPU 10 stores the template indicating the identified area in the RAM 12 as a tracking template, so as to update an eye area in a series of two image frames in chronological order.

摘要翻译： 通过电缆C获取由相机2拾取的图像帧，计算机1的CPU 10识别图像帧中的眼睛区域，并将指示识别的区域的模板存储为跟踪模板。 此外，CPU 10基于跟踪模板识别随后将要获得的图像帧中的眼睛区域，并且计算识别区域的模板和跟踪模板之间的相关值。 此外，当计算出的相关值大于预先存储在HD11中的阈值时，CPU 10将指示识别区域的模板存储在RAM 12中作为跟踪模板，以便更新一系列 两幅图像帧按时间顺序排列。

公开(公告)号：US20050219058A1

公开(公告)日：2005-10-06

申请号：US11135431

申请日：2005-05-24

申请人： Taku Katagiri ,  Tomonobu Takashima ,  Masatoshi Tohno ,  Daisuke Ueno ,  Tadao Omi ,  Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku

发明人： Taku Katagiri ,  Tomonobu Takashima ,  Masatoshi Tohno ,  Daisuke Ueno ,  Tadao Omi ,  Hisao Ito ,  Kozo Baba ,  Akiyoshi Tafuku

IPC分类号： B60K28/04 ,  B60K28/06 ,  G08B21/06 ,  G08B23/00 ,  G08G1/16

CPC分类号： G08B21/06 ,  B60K28/04 ,  B60K28/06 ,  B60W40/09 ,  B60W2520/10 ,  B60W2520/105 ,  B60W2520/125 ,  B60W2530/18

摘要： An alarm system includes a miniature camera (10), a loudspeaker (20), a vehicle speed sensor (30), a longitudinal G sensor (40), a lateral G sensor (50) and an on-vehicle unit (60). A CPU (60a) of the on-vehicle unit (60) executes respective programs (61)-(64) in a ROM (60h), thereby outputting, if a traveling distance during a period for which a driver D continues to take an abnormal behavior exceeds a predetermined value when a traveling state of an automobile is kept stable, an alarm corresponding to a magnitude of the excess over the predetermined value from the loudspeaker (20). The alarm system is thus configured and is therefore capable of exactly outputting the alarm when a danger actually rises.

摘要翻译： 报警系统包括微型照相机（10），扬声器（20），车速传感器（30），纵向G传感器（40），横向G传感器（50）和车载单元（60）。 车载单元（60）的CPU（60a）在ROM（60h）中执行各自的程序（61） - （64），从而如果在驾驶员D继续期间的行驶距离 当汽车的行驶状态保持稳定时，超出预定值的异常行为，与来自扬声器（20）的超过预定值的超出量的警报相对应。 因此，报警系统被配置，并且因此能够在危险实际上升时精确地输出警报。