公开(公告)号：US08855368B2

公开(公告)日：2014-10-07

申请号：US13477612

申请日：2012-05-22

申请人： Yasushi Tanase ,  Takeshi Miura ,  Yasuhiro Todoroki ,  Toshiki Shiino ,  Kensuke Habuka

发明人： Yasushi Tanase ,  Takeshi Miura ,  Yasuhiro Todoroki ,  Toshiki Shiino ,  Kensuke Habuka

IPC分类号： G06K9/00 ,  G06K9/46

CPC分类号： G06K9/4642 ,  G06K9/00684

摘要： The invention provides an image identification device that classifies block images obtained by dividing a target image into predetermined categories, using a separating plane learning of which has been completed in advance for each of the categories. The image identification device includes a target image input unit inputs the target image, a block image generation unit divides the target image into blocks to generate the block images, a feature quantity computing unit computes feature quantities of the block images, and a category determination unit determines whether the block images are classified into one of the categories or not, using the separating plane and coordinate positions corresponding to magnitudes of feature quantities of the block images in a feature quantity space, wherein the feature quantity computing unit uses, as a feature quantity of a given target block image, local feature quantities and a global feature quantity.

摘要翻译： 本发明提供了一种图像识别装置，其将通过将目标图像划分为预定类别而获得的块图像，其中已经为每个类别预先完成了分离平面学习。 图像识别装置包括目标图像输入单元输入目标图像，块图像生成单元将目标图像划分为块以生成块图像，特征量计算单元计算块图像的特征量，以及类别确定单元 使用与特征量空间中的块图像的特征量的大小对应的分离平面和坐标位置来确定块图像是否被分类为其中一个类别，其中特征量计算单元使用特征量计算单元 给定的目标块图像，局部特征量和全局特征量。

公开(公告)号：US08363933B2

公开(公告)日：2013-01-29

申请号：US12462869

申请日：2009-08-11

申请人： Yasuyo Yasuda ,  Tetsuya Hamada ,  Takashi Suzuki ,  Masaki Hiraga ,  Yasushi Tanase ,  Toshiki Shiino

发明人： Yasuyo Yasuda ,  Tetsuya Hamada ,  Takashi Suzuki ,  Masaki Hiraga ,  Yasushi Tanase ,  Toshiki Shiino

IPC分类号： G06K9/00 ,  G06K9/34

CPC分类号： G06K9/00624 ,  G06K9/4642 ,  H04N5/232

摘要： An image identification method for classifying block images of input image data into one of predetermined categories; the method includes the steps of: dividing image data into multiple blocks to produce block images, processing the feature quantity of each block image by their color space information and frequency component, learning separating hyperplanes that indicate boundaries of each category by reading in training data image that have labeled categories for each block and processing image feature quantity for each block of an training data image, and classifying respective block image to a category according to the distance from the separating hyperplane of each category for a newly acquired image to obtain the image feature quantity of block images. An imaging apparatus implementing the image identification method noted above is also disclosed.

摘要翻译： 一种用于将输入图像数据的块图像分类为预定类别之一的图像识别方法; 该方法包括以下步骤：将图像数据划分成多个块以产生块图像，通过其颜色空间信息和频率分量来处理每个块图像的特征量，通过读取训练数据图像来学习分离表示每个类别的边界的超平面 对于每个块具有标记的类别并且针对训练数据图像的每个块处理图像特征量，并且根据与新获取的图像的每个类别的分离超平面的距离将各个块图像分类到类别以获得图像特征 块图像的数量。 还公开了实现上述图像识别方法的成像装置。

公开(公告)号：US20120301015A1

公开(公告)日：2012-11-29

申请号：US13475152

申请日：2012-05-18

申请人： Yasushi TANASE ,  Takeshi MIURA ,  Yasuhiro TODOROKI ,  Toshiki SHIINO ,  Kensuke HABUKA ,  Tetsuya HAMADA ,  Yasuyo YASUDA

发明人： Yasushi TANASE ,  Takeshi MIURA ,  Yasuhiro TODOROKI ,  Toshiki SHIINO ,  Kensuke HABUKA ,  Tetsuya HAMADA ,  Yasuyo YASUDA

IPC分类号： G06K9/62

CPC分类号： G06K9/00684 ,  G06F17/3025 ,  G06K9/4642

摘要： The invention provides an image identification device uses a separating plane to classify block images into the categories. The image identification device includes a target image input unit inputting a target image, a block image generation unit generates block images, a feature quantity computing unit computes feature quantities of the block images, and a category determination unit determines whether the block images are classified into the categories or not. The feature quantity computing unit uses local feature quantities of the block images and a global feature quantity of the target image as a whole, and also in a feature quantity space using features of the block images as coordinate axes, uses coordinate positions of feature quantity vectors optional areas in the feature quantity space to count the block images and causes the global feature quantity to include the number of the block images thus counted.

摘要翻译： 本发明提供了一种图像识别装置，其使用分离平面将块图像分类成类别。 图像识别装置包括输入目标图像的目标图像输入单元，块图像生成单元生成块图像，特征量计算单元计算块图像的特征量，并且类别确定单元确定块图像是否被分类 类别与否。 特征量计算单元使用块图像的局部特征量和目标图像的全局特征量作为整体，并且还使用使用块图像的特征作为坐标轴的特征量空间，使用特征量向量的坐标位置 特征量空间中的可选区域来对块图像进行计数，并使全局特征量包括如此计数的块图像的数量。

公开(公告)号：US20120300982A1

公开(公告)日：2012-11-29

申请号：US13477612

申请日：2012-05-22

申请人： Yasushi TANASE ,  Takeshi Miura ,  Yasuhiro Todoroki ,  Toshiki Shiino ,  Kensuke Habuka

发明人： Yasushi TANASE ,  Takeshi Miura ,  Yasuhiro Todoroki ,  Toshiki Shiino ,  Kensuke Habuka

IPC分类号： G06K9/62

CPC分类号： G06K9/4642 ,  G06K9/00684

摘要： The invention provides an image identification device that classifies block images obtained by dividing a target image into predetermined categories, using a separating plane learning of which has been completed in advance for each of the categories. The image identification device includes a target image input unit inputs the target image, a block image generation unit divides the target image into blocks to generate the block images, a feature quantity computing unit computes feature quantities of the block images, and a category determination unit determines whether the block images are classified into one of the categories or not, using the separating plane and coordinate positions corresponding to magnitudes of feature quantities of the block images in a feature quantity space, wherein the feature quantity computing unit uses, as a feature quantity of a given target block image, local feature quantities and a global feature quantity.

摘要翻译： 本发明提供了一种图像识别装置，其将通过将目标图像划分为预定类别而获得的块图像，其中已经为每个类别预先完成了分离平面学习。 图像识别装置包括目标图像输入单元输入目标图像，块图像生成单元将目标图像划分为块以生成块图像，特征量计算单元计算块图像的特征量，以及类别确定单元 使用与特征量空间中的块图像的特征量的大小对应的分离平面和坐标位置来确定块图像是否被分类为其中一个类别，其中特征量计算单元使用特征量计算单元 给定的目标块图像，局部特征量和全局特征量。

公开(公告)号：US08600162B2

公开(公告)日：2013-12-03

申请号：US13475152

申请日：2012-05-18

申请人： Yasushi Tanase ,  Takeshi Miura ,  Yasuhiro Todoroki ,  Toshiki Shiino ,  Kensuke Habuka ,  Tetsuya Hamada ,  Yasuyo Yasuda

发明人： Yasushi Tanase ,  Takeshi Miura ,  Yasuhiro Todoroki ,  Toshiki Shiino ,  Kensuke Habuka ,  Tetsuya Hamada ,  Yasuyo Yasuda

IPC分类号： G06K9/00

CPC分类号： G06K9/00684 ,  G06F17/3025 ,  G06K9/4642

摘要： The invention provides an image identification device uses a separating plane to classify block images into the categories. The image identification device includes a target image input unit inputting a target image, a block image generation unit generates block images, a feature quantity computing unit computes feature quantities of the block images, and a category determination unit determines whether the block images are classified into the categories or not. The feature quantity computing unit uses local feature quantities of the block images and a global feature quantity of the target image as a whole, and also in a feature quantity space using features of the block images as coordinate axes, uses coordinate positions of feature quantity vectors optional areas in the feature quantity space to count the block images and causes the global feature quantity to include the number of the block images thus counted.

摘要翻译： 本发明提供了一种图像识别装置，其使用分离平面将块图像分类成类别。 图像识别装置包括输入目标图像的目标图像输入单元，块图像生成单元生成块图像，特征量计算单元计算块图像的特征量，并且类别确定单元确定块图像是否被分类 类别与否。 特征量计算单元使用块图像的局部特征量和目标图像的全局特征量作为整体，并且还使用使用块图像的特征作为坐标轴的特征量空间，使用特征量向量的坐标位置 特征量空间中的可选区域来对块图像进行计数，并使全局特征量包括如此计数的块图像的数量。

公开(公告)号：US20100142807A1

公开(公告)日：2010-06-10

申请号：US12462869

申请日：2009-08-11

申请人： Yasuyo Yasuda ,  Tetsuya Hamada ,  Takashi Suzuki ,  Masaki Hiraga ,  Yasushi Tanase ,  Toshiki Shiino

发明人： Yasuyo Yasuda ,  Tetsuya Hamada ,  Takashi Suzuki ,  Masaki Hiraga ,  Yasushi Tanase ,  Toshiki Shiino

IPC分类号： G06K9/34

CPC分类号： G06K9/00624 ,  G06K9/4642 ,  H04N5/232

摘要： An image identification method for classifying block images of input image data into one of the multiple predetermined categories according to feature quantity in each block image; the method includes an image production step of dividing image data into multiple blocks to produce block images, an image feature quantity processing step of processing the feature quantity of each block image by their color space information and frequency component, a separating hyperplane processing step of learning separating hyperplanes that indicate boundaries of each category by reading in training data image that have labeled categories for each block and processing image feature quantity for each block of an training data image, and a category classification step of classifying respective block image to a category according to the distance from the separating hyperplane of each category by executing the block image production step and the image feature quantity processing step for a newly acquired image to obtain the image feature quantity of block images.

摘要翻译： 一种图像识别方法，用于根据每个块图像中的特征量将输入图像数据的块图像分类成多个预定类别中的一个; 该方法包括图像产生步骤，将图像数据分割成多个块以产生块图像;图像特征量处理步骤，通过其颜色空间信息和频率分量来处理每个块图像的特征量;分离超平面处理步骤，学习 通过读取具有每个块的标记类别的训练数据图像和训练数据图像的每个块的处理图像特征量来分离指示每个类别的边界的超平面，以及类别分类步骤，根据 通过对新获取的图像执行块图像生成步骤和图像特征量处理步骤来获得每个类别的分离超平面的距离，以获得块图像的图像特征量。