会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 2. 发明专利
    • Apparatus for detecting face feature point, method of detecting face feature point, and program
    • 用于检测脸部特征点的装置,检测脸部特征点的方法和程序
    • JP2008209969A
    • 2008-09-11
    • JP2007043332
    • 2007-02-23
    • Aisin Seiki Co LtdToyota Central R&D Labs IncToyota Motor Corpアイシン精機株式会社トヨタ自動車株式会社株式会社豊田中央研究所
    • SUZUKI TOMOHARUADACHI ATSUSHIYOSHINAGA NORIHIKOUOZUMI SHIGEYASUNINAGAWA YUJIOGAMI KENICHITAKAHASHI KENTARONAKANISHI SATORUKOJIMA SHINICHISHIRAKI NOBUMASA
    • G06T1/00
    • PROBLEM TO BE SOLVED: To provide an apparatus for detecting face feature point capable of efficiently and securely detecting a feature point of a face such as a nostril or an eye etc. even when the face dose not face a front.
      SOLUTION: The apparatus comprises an image input unit 21 for acquiring a face image, a face center detecting unit 22 for detecting a face center position from the face image, a nostril detecting unit 24 for detecting the nostril from the face image, reference nostril position setting unit 25 for storing a position of the nostril detected by the nostril detecting unit 24 as a reference position when the face center position detected by the face center detecting unit 22 is existing nearly the center of the face image, and a nostril determining unit 26 for determining a nostril candidate whose coordinate of a horizontal direction is nearest to the face center position and whose coordinate of a vertical direction is nearest to the reference nostril position as the nostril from among nostril candidates to be detected by the nostril detecting unit 24 when the face center position detected by the face center detecting unit 22 does not exist at nearly a center within a predetermined range, or when the nostril is detected at the reference position.
      COPYRIGHT: (C)2008,JPO&INPIT
    • 要解决的问题:即使当面部不面向前方时,提供一种用于检测面部特征点的装置,其能够有效且可靠地检测诸如鼻孔或眼睛等的面部的特征点。 解决方案:该装置包括用于获取面部图像的图像输入单元21,用于从面部图像检测面部中心位置的面部中心检测单元22,用于从面部图像检测鼻孔的鼻孔检测单元24, 参考鼻孔位置设定单元25,用于当由面部中心检测单元22检测到的面部中心位置几乎存在于面部图像的中心时,将由鼻孔检测单元24检测到的鼻孔的位置存储为基准位置,并且鼻孔 确定单元26,用于确定由鼻孔检测单元检测的鼻孔候选之间的水平方向的坐标最接近面部中心位置并且其垂直方向的坐标最接近参考鼻孔位置的鼻孔候选, 如图24所示,当由面部中心检测单元22检测到的面部中心位置在预定范围内的近似中心处不存在时, 或者当在参考位置检测到鼻孔时。 版权所有(C)2008,JPO&INPIT
    • 4. 发明专利
    • Eyelid detection apparatus, eyelid detection method, and program
    • 眼睛检测装置,眼睛检测方法和程序
    • JP2008158922A
    • 2008-07-10
    • JP2006349031
    • 2006-12-26
    • Aisin Seiki Co LtdToyota Central R&D Labs IncToyota Motor Corpアイシン精機株式会社トヨタ自動車株式会社株式会社豊田中央研究所
    • ADACHI ATSUSHIASANO TADASHININAGAWA YUJIOGAMI KENICHIKOJIMA SHINICHI
    • G06T1/00
    • A61B3/14A61B3/113A61B5/1176A61B5/18G06K9/0061G06K9/00832
    • PROBLEM TO BE SOLVED: To provide a device for accurately detecting the eyelids even in the case of a red eye phenomenon. SOLUTION: An eyelid detection device 1 is provided with: a camera 2 and an image input part 21 for acquiring a face image; a red eye detection part 25 for detecting a red eye phenomenon by sensing the change of the spatial contrast of an iris part from the face image; and an eyelid candidate detection part 23 and an eyelid decision part 27 for detecting the upper eyelid and the lower eyelid based on the edge as the boundary of contrast from the face image. When the red eye phenomenon is detected by the red eye detection part 25, a red eye removal processing part 26 excludes the upper edge or lower edge of the pupil as the object of red eye phenomenon from the lower eyelid or the upper eyelid. The red eye detection part 25 detects the section of the face image matched with a predetermined contrast pixel pattern or the edge pattern as the pupil being the red eye phenomenon. The predetermined contrast pixel pattern may be configured of a bright region at the central part and a dark region arranged in the periphery. The red eye detection part 25 may detect the section of the face image matched with any of the plurality of different predetermined contrast pixel patterns as the pupil as the red eye phenomenon. COPYRIGHT: (C)2008,JPO&INPIT
    • 要解决的问题:即使在红眼现象的情况下,也提供用于准确地检测眼睑的装置。 眼睑检测装置1设置有摄像机2和用于获取脸部图像的图像输入部21; 用于通过感测来自脸部图像的虹膜部分的空间对比度的变化来检测红眼现象的红眼检测部分25; 以及眼睑候选检测部23和用于基于作为来自面部图像的对比度的边界的边缘检测上眼睑和下眼睑的眼睑判定部27。 当红眼检测部25检测到红眼现象时,红眼消除处理部26从下眼睑或上眼睑排除作为红眼现象的瞳孔的上边缘或下边缘。 红眼检测部25检测与预定的对比像素图案对应的面部图像的部分或作为红眼现象的瞳孔的边缘图案。 预定的对比度像素图案可以由中央部分的亮区域和布置在周边的暗区域构成。 红眼检测部25可以检测与多个不同的预定对比像素图案中的任何一个匹配的面部图像作为瞳孔作为红眼现象。 版权所有(C)2008,JPO&INPIT
    • 5. 发明专利
    • Eye detection device, eye detection method and program
    • 眼睛检测装置,眼睛检测方法和程序
    • JP2008146172A
    • 2008-06-26
    • JP2006329831
    • 2006-12-06
    • Aisin Seiki Co LtdToyota Central R&D Labs IncToyota Motor Corpアイシン精機株式会社トヨタ自動車株式会社株式会社豊田中央研究所
    • ADACHI ATSUSHIISHIGURO HIROSHIOGAWA KENICHIYOSHINAGA NORIHIKOOGAMI KENICHIUOZUMI SHIGEYASUKOJIMA SHINICHINAKANISHI SATORUSHIRAKI NOBUMASA
    • G06T7/00G06T1/00G08G1/16
    • G06K9/00604G06K9/00281
    • PROBLEM TO BE SOLVED: To provide an eye detection device for quickly and surely detecting the eyes from a face image.
      SOLUTION: This eye detection device is provided with an image input part 21 for acquiring a face image; a nostril detection part 24 for detecting nostrils from the face image; an eye retrieval region setting part 25 for setting an eye retrieval region in the face image on the basis of the positions of the nostrils detected by the nostril detection part 24; and a lower eyelid detection part 26 and an upper eyelid detection part 27 for retrieving the eyes under prescribed conditions from the lower part to upper part of the face image in the range of the eye retrieval range set by the eye retrieval region setting part 25. It is desired that this eye detection device includes a mouth detection part 22 for detecting the position of the mouth from the face image, and when detecting a plurality of nostril candidates suitable for the retrieval conditions of the nostrils, the nostril detection part 24 judges the nostril candidate which is closest to the position of the mouth among the plurality of nostril candidates as the nostrils. Furthermore, this eye detection device includes a nostril retrieval region setting part 23 for setting the region where there is high possibility that the nostrils exist in the face image as the nostril retrieval region, and the nostril retrieval region is retrieved under prescribed conditions, and two adjacent dark places are judged as nostrils.
      COPYRIGHT: (C)2008,JPO&INPIT
    • 要解决的问题:提供一种用于从脸部图像快速且可靠地检测眼睛的眼睛检测装置。

      解决方案:该眼睛检测装置设置有用于获取脸部图像的图像输入部21; 用于从面部图像检测鼻孔的鼻孔检测部24; 基于由鼻孔检测部24检测出的鼻孔的位置,设定眼睛检索区域的眼睛检索区域设定部25; 以及下眼睑检测部26和上眼睑检测部27,用于在由眼睛检索区域设定部25设定的眼睛检索范围的范围内,从规定条件下从面部图像的下部到上部取出眼睛。 希望这种眼睛检测装置包括用于从脸部图像检测口部的位置的嘴部检测部分22,并且当检测到适合于鼻孔的检索条件的多个鼻孔候选者时,鼻孔检测部件24判断 鼻孔候选者最接近作为鼻孔的多个鼻孔候选者中的嘴的位置。 此外,这种眼睛检测装置包括鼻孔检索区域设定部23,用于设定面部图像中存在鼻孔的可能性高的区域作为鼻孔检索区域,并且在规定条件下检索鼻孔检索区域, 相邻的黑暗地方被认为是鼻孔。 版权所有(C)2008,JPO&INPIT

    • 6. 发明专利
    • Face feature point detection device and program
    • 表面特征点检测装置和程序
    • JP2010244178A
    • 2010-10-28
    • JP2009090065
    • 2009-04-02
    • Aisin Seiki Co LtdDenso CorpToyota Central R&D Labs IncToyota Motor Corpアイシン精機株式会社トヨタ自動車株式会社株式会社デンソー株式会社豊田中央研究所
    • NAKANISHI SATORUNINOMIYA YOSHIKINAITO TAKASHIKOJIMA SHINICHITAKEMAE KASHUOMI TAKUHIROADACHI ATSUSHIHIRAMAKI TAKASHI
    • G06T1/00G06T7/00
    • G06K9/00281G06T7/73
    • PROBLEM TO BE SOLVED: To accurately detect, as facial feature points, the position of the inner corner of the eye and the position of the tail of the eye, even when the inner corner of the eye and the tail of the eye are concealed due to noise. SOLUTION: A Bezier curve represented by a fixed control point P 3 , indicating a first position of the inner corner of the eye detected from an image, a fixed control point P 4 indicating a first position of the tail of the eye, a control point P 1 (first parameter) corresponding to an upper eyelid position candidate, and a control point P 2 (second parameter) corresponding to a lower eyelid position candidate is defined as a first eyelid shape model. In a second eyelid shape model, wherein P 1 and P 2 are fixed so that a fitting evaluation value λ between the first eyelid shape model and the eyelid shape contained in the image is largest, a control point P 3 (third parameter) indicating a position candidate of the inner corner of the eye and a control point P 4 (fourth parameter), indicating a position candidate for the tail of the eye are changed so that the fitting evaluation value λ becomes greatest; and the P 3 and P 4 are determined as a second position of the inner corner of the eye and a second position of the tail of the eye. COPYRIGHT: (C)2011,JPO&INPIT
    • 要解决的问题:即使当眼睛的内角和眼睛的尾部也能够精确地检测眼睛的内角和眼睛的尾部的位置作为面部特征点 被噪音隐藏。

      解决方案:由固定控制点P 3 表示的贝塞尔曲线,指示从图像检测到的眼睛的内角的第一位置,固定控制点P / SB>表示眼睛的尾部的第一位置,对应于上眼睑位置候选的控制点P 1 (第一参数)和控制点P 2 1 和P 2 被固定,使得包含在图像中的第一眼睑形状模型和眼睑形状之间的拟合评估值λ 指示眼睛内角的位置候选的控制点P 3 (第三参数)和控制点P 4 (第四参数),表示 眼睛尾部的位置候选被改变,使得拟合评估值λ变得最大; 并且将P 3 和P 4 确定为眼睛的内角的第二位置和眼尾的第二位置。 版权所有(C)2011,JPO&INPIT

    • 8. 发明专利
    • Eye position detection device, method and program
    • 眼睛位置检测装置,方法和程序
    • JP2008250811A
    • 2008-10-16
    • JP2007093125
    • 2007-03-30
    • Aisin Seiki Co LtdToyota Central R&D Labs IncToyota Motor Corpアイシン精機株式会社トヨタ自動車株式会社株式会社豊田中央研究所
    • NAKANISHI SATORUNINOMIYA YOSHIKIOGAMI KENICHININAGAWA YUJIUOZUMI SHIGEYASUSUZUKI TOMOHARUADACHI ATSUSHI
    • G06T1/00G06T7/60
    • PROBLEM TO BE SOLVED: To accurately detect an eye position of a subject without being influenced by the state of the subject.
      SOLUTION: A CPU detects an upward horizontal edge and a downward horizontal edge from a face image (step S3), and forms horizontal projection histograms thereof, respectively (step S4). One of peaks of the horizontal projection histogram of the downward horizontal edge is set as an upper eyelid candidate position, and two adjacent peaks of the horizontal projection histogram of the upward horizontal edge which are located at both sides of the upper eyelid candidate position are set as face component lower position candidates (step S5). Further, for each of all set combinations, a test quantity Score of upper and lower eye positions is calculated (step S6), and an eye position (upper eyelid position and lower eyelid position) is determined using the upper eyelid candidate position and face part lower position candidate with the smallest test quantity Score of the calculated test quantities Score (step S7).
      COPYRIGHT: (C)2009,JPO&INPIT
    • 要解决的问题:准确地检测被检体的眼睛位置而不受被检体的状态的影响。 解决方案:CPU从脸部图像检测向上的水平边缘和向下的水平边缘(步骤S3),并分别形成水平投影直方图(步骤S4)。 向下水平边缘的水平投影直方图的一个峰值被设置为上眼睑候选位置,并且设置位于上眼睑候选位置的两侧的向上水平边缘的水平投影直方图的两个相邻峰值 作为面部成分下位置候补(步骤S5)。 此外,对于所有组合中的每一个,计算上眼睛位置和下眼睛位置的测试量分数(步骤S6),并且使用上眼睑候选位置和面部部分确定眼睛位置(上眼睑位置和下眼睑位置) 具有最小测试数量的较低位置候选者计算的测试量分数的分数(步骤S7)。 版权所有(C)2009,JPO&INPIT
    • 9. 发明专利
    • Eye tilt detection apparatus and program
    • 眼睛倾斜检测装置和程序
    • JP2008234209A
    • 2008-10-02
    • JP2007071525
    • 2007-03-19
    • Aisin Seiki Co LtdToyota Central R&D Labs IncToyota Motor Corpアイシン精機株式会社トヨタ自動車株式会社株式会社豊田中央研究所
    • KOJIMA SHINICHIOGAMI KENICHININAGAWA YUJIUOZUMI SHIGEYASUADACHI ATSUSHISUZUKI TOMOHARU
    • G06T1/00G06T7/60
    • PROBLEM TO BE SOLVED: To detect right and left eye tilts individually from a face image.
      SOLUTION: A face image picked up is captured (100, 102). A right eye search area and a left eye search area are set on the face image (103, 104). The right eye image and left eye image are each edge-processed (108). On each of the right eye edge image and left eye edge image, edge feature points of an edge corresponding to a lower eyelid border are detected (110). For each of the right eye edge image and left eye edge image, the detected edge feature points are histogram-projected in each projection direction to calculate a histogram of the edge feature points (112, 116). For each of the right and left eyes, the projection direction of the histogram whose maximum value is the largest is selected, and the angle of the selected projection direction to a reference direction is detected as an eye tilt (114, 118).
      COPYRIGHT: (C)2009,JPO&INPIT
    • 要解决的问题:从脸部图像单独检测左眼和左眼的倾斜。

      解决方案:拍摄的脸部图像被捕获(100,102)。 在脸部图像(103,104)上设置右眼搜索区域和左眼搜索区域。 右眼图像和左眼图像都进行边缘处理(108)。 在每个右眼边缘图像和左眼边缘图像上,检测与下眼睑边缘相对应的边缘的边缘特征点(110)。 对于每个右眼边缘图像和左眼边缘图像,检测到的边缘特征点在每个投影方向上被直方图投影,以计算边缘特征点(112,116)的直方图。 对于右眼和左眼中的每一个,选择最大值最大的直方图的投影方向,并且将所选择的投影方向相对于参考方向的角度检测为眼睛倾斜(114,118)。 版权所有(C)2009,JPO&INPIT