公开(公告)号：US07260268B2

公开(公告)日：2007-08-21

申请号：US10443926

申请日：2003-05-22

申请人： Mitsuyasu Asano ,  Kazuhiko Ueda ,  Takeshi Kubozono ,  Kazuki Yokoyama

发明人： Mitsuyasu Asano ,  Kazuhiko Ueda ,  Takeshi Kubozono ,  Kazuki Yokoyama

IPC分类号： G06K9/40

CPC分类号： G06T7/13

摘要： An input signal is accurately separated into a component in which a change in signal level is sharp and having a large edge, and the other components.A narrow band non-linear filter extracts an edge component in a relatively narrow frequency band from an input signal and outputs the signal. A wide band non-linear filter extracts an edge component in a relatively wide frequency band from the input signal and outputs the signal. A selector outputs a signal composed of the signal where an edge portion is replaced, based on a differentiated value of the signal sent by a differentiator, as the edge component of the input signal. An adder subtracts an output signal from the input signal and outputs the calculated output signal as the edge component of the input signal. This process, may be applied to an image processor.

摘要翻译： 输入信号被准确地分离成其中信号电平的变化是尖锐的并且具有大的边缘的分量，以及其它分量。 窄带非线性滤波器从输入信号中提取相对较窄频带的边缘分量并输出该信号。 宽带非线性滤波器从输入信号中提取相对宽的频带中的边缘分量并输出信号。 选择器输出由边缘部分被替换的信号组成的信号，基于由微分器发送的信号的微分值作为输入信号的边缘分量。 加法器从输入信号中减去输出信号，并输出计算出的输出信号作为输入信号的边沿分量。 该过程可以应用于图像处理器。

公开(公告)号：US07558435B2

公开(公告)日：2009-07-07

申请号：US10805207

申请日：2004-03-22

申请人： Shintaro Okada ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Kazuki Yokoyama

发明人： Shintaro Okada ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Kazuki Yokoyama

IPC分类号： G06K9/40 ,  H04N5/00 ,  H04N5/21

CPC分类号： H04N5/21 ,  H04N5/142

摘要： In each of stages of a nonlinear filter, a successively determined target signal and plural adjacent signals are weighted-averaged with plural sets of different coefficients to calculate plural smoothed signals. Differences in level between the target signal and each of adjacent signals and signals arranged within the predetermined interval from the each of the adjacent signals are compared with a threshold value. According to the comparison, flags are set to the adjacent signal, to an adjacent signal arranged at a symmetrical position, and to adjacent signals arranged beyond the adjacent signals. The intervals of adjacent signals are different among stages. The target signal in level is used instead of the adjacent signals with smoothed signals, which are selected according to the flags. Weighting coefficients may be calculated on the basis of the comparison result, and smoothed signals may be synthesized according to the weighting coefficients.

摘要翻译： 在非线性滤波器的每个级中，连续确定的目标信号和多个相邻信号被加权平均与多组不同系数，以计算多个平滑信号。 将目标信号与布置在与每个相邻信号的预定间隔内的相邻信号和信号中的每一个之间的电平差与阈值进行比较。 根据比较，将标志设置为相邻信号，布置在对称位置的相邻信号以及布置在相邻信号之外的相邻信号。 相邻信号的间隔在不同阶段之间是不同的。 使用电平的目标信号代替具有平滑信号的相邻信号，其根据标志被选择。 可以基于比较结果计算加权系数，并且可以根据加权系数合成平滑信号。

公开(公告)号：US07106346B2

公开(公告)日：2006-09-12

申请号：US10887436

申请日：2004-07-08

申请人： Kazuki Yokoyama ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Tetsuji Inada

发明人： Kazuki Yokoyama ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Tetsuji Inada

IPC分类号： G09G5/00

CPC分类号： G06T5/002 ,  G06T5/20 ,  G06T2207/20012 ,  G06T2207/20192

摘要： An n-filter according to the present invention includes a nonlinear filter, a pattern detector, and a switch. The nonlinear filter maintains a steep edge whose size is larger than a predetermined threshold in fluctuations of pixels constituting an input image signal and, at the same time, smoothes a non-edge portion that does not include the edge. The pattern detector detects a fine edge in the fluctuations of the pixels constituting the input image signal and notifies the switch that the fine edge exists. The switch outputs the input image signal or an image signal output from the nonlinear filter to the subsequent stage in accordance with the notification from the pattern detector.

摘要翻译： 根据本发明的n滤波器包括非线性滤波器，模式检测器和开关。 非线性滤波器在构成输入图像信号的像素的波动中保持尺寸大于预定阈值的陡峭边缘，并且同时平滑不包括边缘的非边缘部分。 图案检测器检测构成输入图像信号的像素的波动中的细边缘，并向切换器通知精细边缘存在。 根据来自图案检测器的通知，开关将输入图像信号或从非线性滤波器输出的图像信号输出到后级。

公开(公告)号：US07773824B2

公开(公告)日：2010-08-10

申请号：US10553085

申请日：2004-04-19

申请人： Shintaro Okada ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Kazuki Yokoyama

发明人： Shintaro Okada ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Kazuki Yokoyama

IPC分类号： G06K9/40

CPC分类号： H04N1/4092 ,  G06T5/004 ,  G06T5/20 ,  H04N5/208

摘要： The present invention relates to a signal processing apparatus and a method, a recording medium, and a program, in which portions except an edge can be smoothed while the edge whose change in pixel value is steep is held correctly. A pixel of attention is determined in step S11, and a neighboring pixel is determined in step S12. In step S13, a difference in pixel values between the pixel of attention and each neighboring pixel is calculated. In step S14, according to a relationship in size between the difference and a threshold value ε., flags are raised for the neighboring pixel and a neighboring pixel which are symmetrical. Furthermore, a flag is also raised for a neighboring pixel away from, in view of the pixel of attention, the symmetrical neighboring pixel for which the flag is raised. In step S15, 7-pixel taps centered around the pixel of attention are averaged by weight. However, with respect to the neighboring pixel for which the flag is raised, the pixel value is replaced by that of the pixel of attention C, and is calculated. The present invention can be applied to a video camera, a television receiver, etc.

摘要翻译： 信号处理设备和方法，记录介质和程序，其中除边缘以外的部分可被平滑化，同时像素值陡峭的边缘被正确地保持。 在步骤S11中确定关注像素，并且在步骤S12中确定相邻像素。 在步骤S13中，计算注意像素和每个相邻像素之间的像素值的差异。 在步骤S14中，根据差值和阈值之间的大小关系，对于相邻像素和对称的相邻像素，标志被升高。 此外，鉴于注意的像素，标记升高的对称相邻像素的标志也被提升为相邻像素。 在步骤S15中，以注意像素为中心的7像素抽头通过重量进行平均。 然而，对于标志升高的相邻像素，像素值被注意C的像素值代替，并被计算。 本发明可应用于摄像机，电视接收机等

公开(公告)号：US20050035977A1

公开(公告)日：2005-02-17

申请号：US10887436

申请日：2004-07-08

申请人： Kazuki Yokoyama ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Tetsuji Inada

发明人： Kazuki Yokoyama ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Tetsuji Inada

IPC分类号： G06T5/20 ,  G06T5/00 ,  H04N1/409 ,  H04N5/205 ,  H04N5/235 ,  G09G5/00

CPC分类号： G06T5/002 ,  G06T5/20 ,  G06T2207/20012 ,  G06T2207/20192

摘要： An n-filter according to the present invention includes a nonlinear filter, a pattern detector, and a switch. The nonlinear filter maintains a steep edge whose size is larger than a predetermined threshold in fluctuations of pixels constituting an input image signal and, at the same time, smoothes a non-edge portion that does not include the edge. The pattern detector detects a fine edge in the fluctuations of the pixels constituting the input image signal and notifies the switch that the fine edge exists. The switch outputs the input image signal or an image signal output from the nonlinear filter to the subsequent stage in accordance with the notification from the pattern detector.

摘要翻译： 根据本发明的n滤波器包括非线性滤波器，模式检测器和开关。 非线性滤波器在构成输入图像信号的像素的波动中保持尺寸大于预定阈值的陡峭边缘，并且同时平滑不包括边缘的非边缘部分。 图案检测器检测构成输入图像信号的像素的波动中的细边缘，并向切换器通知精细边缘存在。 根据来自图案检测器的通知，开关将输入图像信号或从非线性滤波器输出的图像信号输出到后级。

公开(公告)号：US20060215927A1

公开(公告)日：2006-09-28

申请号：US10553085

申请日：2004-04-19

申请人： Shintaro Okada ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Kazuhiko Yokoyama

发明人： Shintaro Okada ,  Kazuhiko Ueda ,  Mitsuyasu Asano ,  Takeshi Kubozono ,  Kazuhiko Yokoyama

IPC分类号： G06K9/40

CPC分类号： H04N1/4092 ,  G06T5/004 ,  G06T5/20 ,  H04N5/208

摘要： The present invention relates to a signal processing apparatus and a method, a recording medium, and a program, in which portions except an edge can be smoothed while the edge whose change in pixel value is steep is held correctly. A pixel of attention is determined in step S11, and a neighbouring pixel is determined in step S12. In step S13, a difference in pixel values between the pixel of attention and each neighbouring pixel is calculated. In step S14, according to a relationship in size between the difference and a threshold value ε., flags are raised for the neighbouring pixel and a neighbouring pixel which are symmetrical. Furthermore, a flag is also raised for a neighbouring pixel away from, in view of the pixel of attention, the symmetrical neighbouring pixel for which the flag is raised. In step S15, 7-pixel taps centered around the pixel of attention are averaged by weight. However, with respect to the neighbouring pixel for which the flag is raised, the pixel value is replaced by that of the pixel of attention C, and is calculated. The present invention can be applied to a video camera, a television receiver, etc.

摘要翻译： 信号处理设备和方法，记录介质和程序，其中除边缘以外的部分可被平滑化，同时像素值陡峭的边缘被正确地保持。 在步骤S11中确定关注像素，并且在步骤S12中确定相邻像素。 在步骤S13中，计算注意像素与每个相邻像素之间的像素值的差异。 在步骤S14中，根据差异和阈值ε之间的大小关系，对于相邻像素和对称的相邻像素而升高标志。 此外，鉴于注意的像素，标记升高的对称相邻像素的标志也被提升为相邻像素。 在步骤S15中，以注意的像素为中心的7像素抽头通过重量进行平均。 然而，对于标志升高的相邻像素，像素值被注意C的像素值代替，并被计算。 本发明可应用于摄像机，电视接收机等

公开(公告)号：US06525770B1

公开(公告)日：2003-02-25

申请号：US09015786

申请日：1998-01-29

申请人： Kuzuhiko Ueda ,  Takeshi Kubozono

发明人： Kuzuhiko Ueda ,  Takeshi Kubozono

IPC分类号： H04N5335

CPC分类号： H04N5/23261 ,  H04N5/23238 ,  H04N5/23248 ,  H04N5/23264 ,  H04N5/2625 ,  H04N5/37213 ,  H04N5/3728 ,  H04N9/045

摘要： The initial value T0 of the period of a four-phase driving signal for driving a CCD of an image pickup device is calculated according to the equation T0=(L×p)/(f×v). Wherein, L represents the distance from the CCD to a product (object), if represents the focus length of the optical system used between the CCD and product, v represents the transferring speed of the product, and p represents the distance interval of photodiodes of the CCD. Afterward, the period T of the driving signal is adjusted properly depending on the pickup position of the product. Because electric charges corresponding to the same portion of the image of the product are accumulated in the same potential well formed by the driving signal by generating the driving signal as described above, the image of the product is picked up with suppressed blurring regardless of the speed of the product. Blurring is prevented during image pickup of an transferring object by the method described above.

摘要翻译： 根据等式T0 =（Lxp）/（fxv）计算用于驱动摄像装置的CCD的四相驱动信号的周期的初始值T0。 其中，L表示从CCD到产品（物体）的距离，如果表示CCD和产品之间使用的光学系统的焦距长度，v表示产品的传送速度，p表示光电二极管的距离间隔 CCD。 之后，根据产品的拾取位置适当调整驱动信号的周期T. 由于与产品的图像的相同部分对应的电荷通过产生如上所述的驱动信号而被积累在由驱动信号形成的相同的势阱中，因此无论速度如何，都以抑制的模糊来拾取产品的图像 的产品。 通过上述方法在转印对象的图像拾取期间防止模糊。

公开(公告)号：US06714592B1

公开(公告)日：2004-03-30

申请号：US09711331

申请日：2000-11-13

申请人： Yu Liu ,  Naofumi Yanagihara ,  Kazushi Sato ,  Takeshi Kubozono ,  Shintaro Okada

发明人： Yu Liu ,  Naofumi Yanagihara ,  Kazushi Sato ,  Takeshi Kubozono ,  Shintaro Okada

IPC分类号： H04N712

CPC分类号： H04N19/40 ,  H04N19/124 ,  H04N19/149 ,  H04N19/152 ,  H04N19/177 ,  H04N19/61

摘要： A picture information conversion method and apparatus in which, if the compressed picture information with a variable or non-recognized information volume is input, optimum compression processing on the compressed picture information is performed to prevent underflow and/or overflow to suppress deterioration of the picture quality. A compressed information analysis device 3 analyzes the structure of the picture type of the GOP of the compressed picture information input to the FF buffer 2. A bit allocation device 5 generates a pseudo GOP of the input compressed picture information, based on N frames stored in the FF buffer 2 and on the analysis result information stored in the information buffer 4, and allocates a pre-set code volume to respective pictures of the generated GOP so that the output compressed picture information will be at a second bitrate.

摘要翻译： 一种图像信息转换方法和装置，其中如果输入具有可变或未识别的信息量的压缩图像信息，则执行对压缩图像信息的最佳压缩处理以防止下溢和/或溢出以抑制图像的劣化 质量。 压缩信息分析装置3分析输入到FF缓冲器2的压缩图像信息的GOP的图像类型的结构。比特分配装置5基于存储在N帧中的N个帧来生成输入的压缩图像信息的伪GOP FF缓冲器2和存储在信息缓冲器4中的分析结果信息，并且将所预定的代码量分配给生成的GOP的各个图像，使得输出的压缩图像信息将处于第二位速率。

公开(公告)号：US06658157B1

公开(公告)日：2003-12-02

申请号：US09605035

申请日：2000-06-27

申请人： Kazushi Satoh ,  Takeshi Kubozono ,  Shintaro Okada ,  Yu Liu ,  Naofumi Yanagihara

发明人： Kazushi Satoh ,  Takeshi Kubozono ,  Shintaro Okada ,  Yu Liu ,  Naofumi Yanagihara

IPC分类号： G06K936

CPC分类号： H04N19/00 ,  G06F17/147 ,  H04N19/122 ,  H04N19/132 ,  H04N19/146 ,  H04N19/152 ,  H04N19/18 ,  H04N19/40 ,  H04N19/61 ,  H04N19/70

摘要： A technique is disclosed for reducing the bit rate of a bit stream coded into a compressed form according to the MPEG-2 standard, without causing significant degradation in image quality of a decoded image. An inverse DCT converter dequantizes a MPEG-2 bit stream and outputs resultant discrete cosine transform coefficients for each DCT block. A band limiter replaces the values of high-frequency components of the discrete cosine transform coefficients in the horizontal direction, with 0. A quantizer quantizes the discrete cosine transform coefficients using a quantization step value different from the original value.

摘要翻译： 公开了一种用于根据MPEG-2标准降低编码成压缩形式的比特流的比特率的技术，而不会导致解码图像的图像质量的显着劣化。 逆DCT转换器对MPEG-2比特流进行去量化，并输出每个DCT块的合成离散余弦变换系数。 频带限制器用水平方向将离散余弦变换系数的高频分量的值替换为0.量化器使用不同于原始值的量化步长值来量化离散余弦变换系数。