-
51.发明公开
公开(公告)号：EP0450521A2
公开(公告)日：1991-10-09
申请号：EP91105033.4
申请日：1991-03-28
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
发明人： Takatori, Sunao , Yamamoto, Makoto
IPC分类号： G06F15/80
CPC分类号： G06N3/082
摘要： A learning method for a data processing system comprising an input layer and output layer which comprises a plurality of neurons each of which outputs an predetermined data after igniting according to the predetermined processing results performed onto an input data, and an middle layer, arranged between the input and output layer, the middle layer comprising a plurality of neurons each of which is connected to each neuron of the input and output layer; characterized in the following steps: the ignition patterns of the input layer and the output layer are determined artificially according to a plurality of inputs and outputs; weights of synapses of the middle layer and the output layer are increased so as to obtain the tendency that the ignition pattern of middle layer becomes the nearest approximation to the ignition patterns of input layer and output layer according to each input and output; the same processings as above are performed with respect to all inputs and outputs.
摘要翻译： 一种用于数据处理系统的学习方法，包括输入层和输出层，其包括多个神经元，每个神经元根据对输入数据执行的预定处理结果在点燃之后输出预定数据，以及中间层，布置在 输入和输出层，中间层包括多个神经元，每个神经元连接到输入和输出层的每个神经元; 其特征在于以下步骤：根据多个输入和输出人工确定输入层和输出层的点火模式; 中间层和输出层的突触重量增加，以获得根据每个输入和输出中间层的点火模式与输入层和输出层的点火模式近似近似的趋势; 对于所有输入和输出执行与上述相同的处理。
-
公开(公告)号：EP0405174A3
公开(公告)日：1991-08-07
申请号：EP90110384
申请日：1990-05-31
申请人： YOZAN INC.
CPC分类号： G06N3/08
-
公开(公告)号：EP0343644A3
公开(公告)日：1991-07-03
申请号：EP89109408.8
申请日：1989-05-24
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
IPC分类号： G06T5/00
CPC分类号： G06T5/20
摘要： An image processing method for density-weighted distance transformation of 4-neighborhoods or 8-neighborhoods, comprising steps of:
holding the original image;
generating an image of satisfactory high density (hereinafter it is called the highest density) when the density of background pixels are lower than pixels of a configuration;
in at the 4-neighborhoods processing, giving each center pixel the lowest density among 4-neighborhoods;
adding the image generated by aid giving the lowest density to the original frame memory with the limit of its highest density;
repeating above two processing until a process result agree with the previous process result;
in the 8-neighborhoods processing, giving each center pixel the lowest density among 8-neighborhoods;
adding the image generated by the giving lowest density to the original frame memory with the limit of its highest density;
repeating above two processing until a process result agree with the previous result.摘要翻译： 一种用于4邻域或8邻域的密度加权距离变换的图像处理方法，包括以下步骤：保持原始图像; 当背景像素的密度低于配置的像素时，产生令人满意的高密度（在下文中称为最高密度）的图像; 在4邻域处理中，给每个中心像素在4邻域中密度最低; 将通过给予最低密度产生的图像添加到具有其最高密度极限的原始帧存储器中; 重复上述两个处理，直到处理结果与先前处理结果一致; 在8邻域处理中，给每个中心像素8个邻域中的最低密度; 将具有最低密度的图像与其最高密度的极限相加添加到原始帧存储器; 重复上述两个处理直到处理结果与先前的结果一致。
-
公开(公告)号：EP0429924A2
公开(公告)日：1991-06-05
申请号：EP90121501.2
申请日：1990-11-09
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
发明人： Takatori, Sunao, c/o Yozan Inc. , Kumagai, Ryohei, c/o Yozan Inc. , Matsumoto, Koji, c/o Yozan Inc. , Yamamoto, Makoto, c/o Yozan Inc.
CPC分类号： G10L15/04
摘要： An acoustic recognition method including step of phoneme segmentation characterized in that: the phoneme segmentation step comprises a step of concavity calculation step for calculating depth of concavities of acoustic power sequence so as to find an acoustic boundary.
摘要翻译： 一种声音识别方法，包括音素分割步骤，其特征在于：音素分割步骤包括凹陷计算步骤的步骤，用于计算声功率序列的凹陷深度，以便找到声学边界。
-
55.发明公开Neuronal data processing system and adaptation method therefor 失效神经元Datenverarbeitungssystem和Anpassungsverfahren。
公开(公告)号：EP0405174A2
公开(公告)日：1991-01-02
申请号：EP90110384.6
申请日：1990-05-31
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
IPC分类号： G06F15/80
CPC分类号： G06N3/08
摘要： An adaptation method for a data processing system comprising a plurality of neurons each of which outputs an output according to a comparison between a sum of multiplied inputs by weights and a threshold, characterized in the following steps:
The threshold of the neuron which has generated significant output at certain point of time is compulsorily increased to a maximal value;
The weight of the neuron is adapted for a constant value of the inputs; and the threshold is decreased to a value at the point.摘要翻译： 一种用于数据处理系统的适应方法，包括多个神经元，每个神经元根据权重和相乘输入之和之间的比较来输出输出，其特征在于以下步骤：产生显着的神经元的阈值 在某个时间点的输出强制增加到最大值; 神经元的权重适用于输入的恒定值; 并且阈值减小到该点处的值。
-
公开(公告)号：EP0337325A3
公开(公告)日：1990-12-19
申请号：EP89106204.4
申请日：1989-04-07
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
摘要： An image processing system for finding a local minimum in a histogram of an image in order to define a threshold comprising steps of; Generating a table of density according to the order of number of pixels; Assuming that the density of maximal number of pixels of the table is one local maximum of higher or lower density; Searching another local maximum and the local minimum based on the assumption; Assuming that the density of maximal number of pixels of the table is another local maximum of lower or higher density; Searching another local maximum and the local minimum based on the second assumption; and Determining true local maximums and the local minimum reviewing the first and second searching.
摘要翻译： 一种用于在图像的直方图中找到局部最小值以便定义阈值的图像处理系统，包括以下步骤： 根据像素数的顺序生成密度表; 假设表的最大像素数的密度是较高或较低密度的一个局部最大值; 根据假设搜索另一个本地最大值和当地最小值; 假设表的最大像素数的密度是较低或较高密度的另一局部最大值; 根据第二个假设搜索另一个局部最大值和局部最小值; 并确定真正的本地最大值和局部最小值来检查第一和第二搜索。
-
公开(公告)号：EP0381794A1
公开(公告)日：1990-08-16
申请号：EP89102267.5
申请日：1989-02-09
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
IPC分类号： G06T9/20
CPC分类号： G06T9/20
摘要： The present invention has an object to provide a method for registering image data, effective for high speed image inputting and reappearing.
According to the method of the present invention, the end points of each line included in a configuration are extracted and registered.摘要翻译： 本发明的目的是提供一种用于登记图像数据的方法，其有效用于高速图像输入和再现。 根据本发明的方法，提取并登记构成中包含的各行的终点。
-
公开(公告)号：EP0380720A1
公开(公告)日：1990-08-08
申请号：EP89101564.6
申请日：1989-01-30
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
CPC分类号： G06T5/20 , H03H17/0266
摘要： Image processing method for smoothing without dulling edge.
Convolution areas of predetermined size are successively read out, pixels in each convolution area are divided into a plurality of groups, the median value of each group is calculated, the median value of the median values of the groups is calculated and the last median value is defined as a representative value of the convolution.摘要翻译： 用于平滑无图像边缘的图像处理方法。 连续地读出预定尺寸的卷积区域，将各卷积区域中的像素分成多组，计算各组的中值，计算组中值的中位值，最后的中值为 定义为卷积的代表值。
-
公开(公告)号：EP0341701A2
公开(公告)日：1989-11-15
申请号：EP89108437.8
申请日：1989-05-10
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
IPC分类号： G06K9/48
CPC分类号： G06K9/48
摘要： An image processing method for finding concavity of a configuration comprising steps;
arranging boundary pixels of the configuration according to chain order with indicating top points on convex hull of the configuration; and
determining that a concavity is defined by the boundary pixels when a boundary pixel of a chain code is found, which chain code indicates a concavity.摘要翻译： 一种用于发现包括以下步骤的配置的凹面的图像处理方法： 根据链顺序排列配置的边界像素，在配置的凸包上指示顶点; 并且当发现链码的边界像素时，确定凹部由边界像素定义，哪个链码表示凹面。
-
公开(公告)号：EP0337325A2
公开(公告)日：1989-10-18
申请号：EP89106204.4
申请日：1989-04-07
申请人： YOZAN INC. , SHARP KABUSHIKI KAISHA
摘要： An image processing system for finding a local minimum in a histogram of an image in order to define a threshold comprising steps of;
Generating a table of density according to the order of number of pixels;
Assuming that the density of maximal number of pixels of the table is one local maximum of higher or lower density;
Searching another local maximum and the local minimum based on the assumption;
Assuming that the density of maximal number of pixels of the table is another local maximum of lower or higher density;
Searching another local maximum and the local minimum based on the second assumption; and
Determining true local maximums and the local minimum reviewing the first and second searching.摘要翻译： 一种图像处理系统，用于在图像的直方图中找到局部最小值以便定义阈值，该图像处理系统包括以下步骤： 根据像素数量的顺序生成密度表; 假设表格的最大像素数密度是较高或较低密度的一个局部最大值; 根据假设搜索另一个局部最大值和局部最小值; 假设表格的最大像素数密度是较低或较高密度的另一局部最大值; 根据第二个假设搜索另一个局部最大值和局部最小值; 以及确定真正的局部最大值和局部最小值，审查第一次和第二次搜索。
-
-
-
-
-
-
-
-
-
搜索结果
343 条记录 (耗时 0.015 秒)
