公开(公告)号：US07065138B2

公开(公告)日：2006-06-20

申请号：US11078067

申请日：2005-03-10

申请人： Takuya Kitamura ,  Akira Sugiyama ,  Haruo Togashi

发明人： Takuya Kitamura ,  Akira Sugiyama ,  Haruo Togashi

IPC分类号： H04N7/12

CPC分类号： H04N19/14 ,  H04N19/126 ,  H04N19/149 ,  H04N19/15 ,  H04N19/176 ,  H04N19/60

摘要： A residue code amount is calculated. A threshold act_thrsh corresponding to the residue code amount is obtained. A variable mb that designates a macro block is initialized to zero. The variable mb is compared with the total number of macro blocks. When the variable mb matches the total number of macro blocks, the process is completed. Otherwise, the rank of the activity information of the macro block designated by the variable mb is compared with the rank of the threshold act_thrsh. When the rank of the activity information is lower than the rank of the threshold act_thrsh, since there is a possibility of which q_scale can be decreased, q_scale is calculated with the activity information of the macro block mb and min_qq. q_scale of the macro block is subtracted from the total code amount. The subtracted resultant data is stored in a variable tmp. Thereafter, the variable tmp is compared with a target code amount GEN_TGT. When the variable tmp is equal to or smaller than the target code amount GEN_TGT, since the quantizing step can be fined down, the quantizing step of the macro block mb is decreased by one.

摘要翻译： 计算残留代码量。 获得与残留码量对应的阈值act_thrsh。 指定宏块的变量mb被初始化为零。 将变量mb与宏块的总数进行比较。 当变量mb匹配宏块的总数时，该过程完成。 否则，将由变量mb指定的宏块的活动信息的等级与阈值act_thrsh的等级进行比较。 当活动信息的等级低于阈值act_thrsh的等级时，由于存在可以减少q_scale的可能性，所以使用宏块mb和min_qq的活动信息来计算q_scale。 从整个代码量中减去宏块的q_scale。 减去的结果数据存储在变量tmp中。 此后，将变量tmp与目标代码量GEN_TGT进行比较。 当变量tmp等于或小于目标代码量GEN_TGT时，由于量化步骤可以被精细化，所以宏块mb的量化步骤减少一。

公开(公告)号：US06970938B2

公开(公告)日：2005-11-29

申请号：US10032917

申请日：2001-12-27

申请人： Akira Sugiyama ,  Haruo Togashi ,  Shin Todo ,  Hideyuki Matsumoto

发明人： Akira Sugiyama ,  Haruo Togashi ,  Shin Todo ,  Hideyuki Matsumoto

IPC分类号： H03M7/40 ,  H04N19/00 ,  H04N19/423 ,  H04N19/44 ,  H04N19/46 ,  H04N19/70 ,  H04N19/89 ,  H04N19/895 ,  G06F15/16

CPC分类号： H04B1/715 ,  H04B2001/7154 ,  H04N19/89 ,  H04N19/895

摘要： A detection circuit detects a slice start code from an input stream. Based on a result of the detection and a frame pulse, a reference macroblock (MB) address generated by a timing generator is sent to an address comparator. The input stream which has been phase-adjusted by a delay circuit is supplied to a variable length decoder (VLD) for decoding a variable length code to detect an MB address. The address comparator compares the reference MB address with the MB address to check the continuity of the MB address. If it is discontinuous, the stream output from the VLD is temporarily interrupted, and a selector selects a replacement data generating circuit in response to a control signal output from the address comparator. The input stream is replaced at a discontinuous macroblock portion with macroblock data which has a correct macroblock address and which is prepared in advance by the replacement data generating circuit. This provides stable processing of macroblock discontinuities when I-pictures in an MPEG stream are input.

摘要翻译： 检测电路从输入流检测切片起始码。 基于检测结果和帧脉冲，由定时发生器产生的参考宏块（MB）地址被发送到地址比较器。 通过延迟电路进行相位调整的输入流被提供给可变长度解码器（VLD），用于解码可变长度代码以检测MB地址。 地址比较器将参考MB地址与MB地址进行比较，以检查MB地址的连续性。 如果不连续，则从VLD输出的流暂时中断，并且选择器响应于从地址比较器输出的控制信号选择替换数据产生电路。 在具有正确的宏块地址的宏块数据的不连续宏块部分替换输入流，并由替换数据产生电路预先准备。 当输入MPEG流中的I图像时，这提供了宏块不连续性的稳定处理。

公开(公告)号：US06654544B1

公开(公告)日：2003-11-25

申请号：US09436308

申请日：1999-11-08

申请人： Takao Suzuki ,  Haruo Togashi ,  Satoshi Miyazawa

发明人： Takao Suzuki ,  Haruo Togashi ,  Satoshi Miyazawa

IPC分类号： H04N5917

CPC分类号： H04N9/7921 ,  H04N5/765 ,  H04N5/78266 ,  H04N5/783 ,  H04N9/7925 ,  H04N9/8042 ,  H04N9/8047 ,  H04N9/8063 ,  H04N9/8205 ,  H04N9/888 ,  H04N21/236 ,  H04N21/4334 ,  H04N21/434 ,  H04N21/4402

摘要： The present invention is a video data recording apparatus, a video data recording method, a video data reproducing apparatus, a video data reproducing method, a video data recording and reproducing apparatus, and a video data recording and reproducing method that can be interfaced using an MPEG2 elementary stream and that allow an input elementary stream to be recorded to a record medium and reproduced therefrom. According to the present invention, a recording side stream converting means is disposed. The recording side stream converting means converts an input elementary stream into a device stream that is a stream suitable for a record medium. A recording means performs a recording process for the converted device stream and records the record data to the record medium. A reproducing means reproduces record data recorded on the record medium as a device stream. According to the present invention, a reproducing side stream converting means is disposed. The reproducing side stream converting means converts the restored device stream into an elementary stream.

摘要翻译： 本发明是一种视频数据记录装置，视频数据记录方法，视频数据再现装置，视频数据再现方法，视频数据记录和再现装置以及视频数据记录和再现方法，其可以使用 MPEG2基本流，并且允许将输入基本流记录到记录介质并从其再现。根据本发明，设置记录侧流转换装置。 记录侧流转换装置将输入的基本流转换成适合于记录介质的流的设备流。 记录装置对转换的装置流执行记录处理，并将记录数据记录到记录介质上。再现装置将记录在记录介质上的记录数据作为装置流再现。 根据本发明，设置再现侧流转换装置。 再现侧流转换装置将恢复的设备流转换为基本流。

公开(公告)号：US08077932B2

公开(公告)日：2011-12-13

申请号：US11950929

申请日：2007-12-05

申请人： Haruo Togashi

发明人： Haruo Togashi

IPC分类号： G06K9/00

CPC分类号： G06K9/4619 ,  G06K9/00281

摘要： In a face recognition apparatus adapted to recognize an input face image on the basis of one or more registered face images, similarity between an input face image and a registered face image is determined based on a Gabor jet calculated by performing Gabor filtering using Gabor filters defined by a Gaussian function representing a window and sine and cosine functions representing a frequency response for each of predetermined frequency values and for each of predetermined angles of the response function. Values of the filter window are calculated in advance based on the Gaussian function and stored in a filter window data ROM. Values of the response function are calculated in advance for respective angles based on the sine and cosine functions and stored in sine data ROMs and cosine data ROMs. In the Gabor filtering process, coefficients of the Gabor filters are determined from values read from these ROMs.

摘要翻译： 在适于基于一个或多个登记的面部图像来识别输入面部图像的面部识别装置中，基于通过使用定义的Gabor滤波器执行Gabor滤波而计算的Gabor jet来确定输入面部图像和注册面部图像之间的相似度 通过高斯函数表示窗口和表示针对每个预定频率值的频率响应的正弦和余弦函数以及响应函数的每个预定角度。 基于高斯函数预先计算滤波器窗口的值并存储在滤波器窗口数据ROM中。 基于正弦和余弦函数预先计算响应函数的值，并存储在正弦数据ROM和余弦数据ROM中。 在Gabor滤波过程中，Gabor滤波器的系数由从这些ROM读取的值确定。

公开(公告)号：US07916994B2

公开(公告)日：2011-03-29

申请号：US10219409

申请日：2002-08-15

申请人： Shin Todo ,  Haruo Togashi ,  Akira Sugiyama ,  Hideyuki Matsumoto

发明人： Shin Todo ,  Haruo Togashi ,  Akira Sugiyama ,  Hideyuki Matsumoto

IPC分类号： H04N5/91

CPC分类号： H04N19/70 ,  H04N5/78263 ,  H04N9/7921 ,  H04N9/8042

摘要： When a stream encoded with a variable length code is output, a stream containing a syntax error and a format violation is prevented from being output.A checker checks encoding parameter for a stream that is reproduced from a magnetic tape and a stream that is output through an EE path and determines an occurrence of a syntax error or a format violation. When the output stream contains it, a switch circuit stops outputting the stream in the unit of a GOP or a picture and adds a code representing the end to the immediately preceding stream. The checker continuously checks the stream. When the output stream does not contain an error, etc, the output of the stream is resumed from the beginning of a GOP. At that point, flag representing that there is a picture that has not been used due to the stop of the stream is added.

摘要翻译： 当输出用可变长度代码编码的流时，阻止包含语法错误和格式违例的流被输出。 检查器检查从磁带再现的流的编码参数和通过EE路径输出的流，并确定语法错误或格式违规的发生。 当输出流包含它时，开关电路停止以GOP或图像为单位输出流，并将代表结束的代码添加到紧接在前的流。 检查器不断检查流。 当输出流不包含错误等时，从GOP的开始恢复流的输出。 此时，表示存在由于流停止而没有使用的图片的标志。

公开(公告)号：US07145949B2

公开(公告)日：2006-12-05

申请号：US11077642

申请日：2005-03-10

申请人： Takuya Kitamura ,  Akira Sugiyama ,  Haruo Togashi

发明人： Takuya Kitamura ,  Akira Sugiyama ,  Haruo Togashi

IPC分类号： H04N7/12

CPC分类号： H04N19/14 ,  H04N19/126 ,  H04N19/149 ,  H04N19/15 ,  H04N19/176 ,  H04N19/60

摘要： A residue code amount is calculated. A threshold act_thrsh corresponding to the residue code amount is obtained. A variable mb that designates a macro block is initialized to zero. The variable mb is compared with the total number of macro blocks. When the variable mb matches the total number of macro blocks, the process is completed. Otherwise, the rank of the activity information of the macro block designated by the variable mb is compared with the rank of the threshold act_thrsh. When the rank of the activity information is lower than the rank of the threshold act_thrsh, since there is a possibility of which q_scale can be decreased, q_scale is calculated with the activity information of the macro block mb and min_qq. q_scale of the macro block is subtracted from the total code amount. The subtracted resultant data is stored in a variable tmp. Thereafter, the variable tmp is compared with a target code amount GEN_TGT. When the variable tmp is equal to or smaller than the target code amount GEN_TGT, since the quantizing step can be fined down, the quantizing step of the macro block mb is decreased by one.

公开(公告)号：US20050175094A1

公开(公告)日：2005-08-11

申请号：US11077642

申请日：2005-03-10

申请人： Takuya Kitamura ,  Akira Sugiyama ,  Haruo Togashi

发明人： Takuya Kitamura ,  Akira Sugiyama ,  Haruo Togashi

IPC分类号： H04N19/60 ,  H04N19/00 ,  H04N19/126 ,  H04N19/134 ,  H04N19/14 ,  H04N19/15 ,  H04N19/172 ,  H04N19/186 ,  H04N19/189 ,  H04N19/192 ,  H04N19/196 ,  H04N19/423 ,  H04N19/625 ,  H04N19/63 ,  H04N19/91 ,  H04N7/12

CPC分类号： H04N19/14 ,  H04N19/126 ,  H04N19/149 ,  H04N19/15 ,  H04N19/176 ,  H04N19/60

摘要： A residue code amount is calculated. A threshold act_thrsh corresponding to the residue code amount is obtained. A variable mb that designates a macro block is initialized to zero. The variable mb is compared with the total number of macro blocks. When the variable mb matches the total number of macro blocks, the process is completed. Otherwise, the rank of the activity information of the macro block designated by the variable mb is compared with the rank of the threshold act_thrsh. When the rank of the activity information is lower than the rank of the threshold act_thrsh, since there is a possibility of which q_scale can be decreased, q_scale is calculated with the activity information of the macro block mb and min_qq. q_scale of the macro block is subtracted from the total code amount. The subtracted resultant data is stored in a variable tmp. Thereafter, the variable tmp is compared with a target code amount GEN_TGT. When the variable tmp is equal to or smaller than the target code amount GEN_TGT, since the quantizing step can be fined down, the quantizing step of the macro block mb is decreased by one.

公开(公告)号：US06721492B1

公开(公告)日：2004-04-13

申请号：US09418265

申请日：1999-10-14

申请人： Haruo Togashi

发明人： Haruo Togashi

IPC分类号： H04N726

CPC分类号： G11B27/3027 ,  G11B20/10 ,  G11B20/1252 ,  H04N5/783 ,  H04N9/7921 ,  H04N9/8042 ,  H04N9/8063 ,  H04N19/129 ,  H04N19/18 ,  H04N19/625 ,  H04N19/88

摘要： In order to obtain a high-quality image when an encoded stream which is encoded using MPEG is played back at variable speed, a stream conversion apparatus is provided which converts an arrangement of DCT coefficients of a source encoded stream into an arrangement suitable for variable-speed playback. Specifically, a header part is separated from an elementary stream of MPEG, and DCT coefficients arranged from a low order of a spatial frequency to a high order are extracted by zigzag scan at intervals of a plurality of DCT blocks which form a macro block. These DCT coefficients are rearranged from a low order to a high order across a plurality of DCT blocks. The DCT coefficients rearranged in this manner are packed into a stream, and the separated header part is added, forming a converted elementary stream. This converted elementary stream is recorded. During high-speed playback, even if there is an error in the middle of the macro block, since decoding is possible using a low-order DCT coefficient, high-quality image video data can be obtained.

摘要翻译： 为了在使用MPEG编码的编码流以可变速度进行再现的情况下获得高品质图像，提供了一种流转换装置，其将源编码流的DCT系数的排列转换成适于可变频率的装置， 速度播放。 具体地说，将标题部分与MPEG的基本流分开，并且从空间频率的低阶到高阶排列的DCT系数以形成宏块的多个DCT块的间隔的Z字形扫描提取。 这些DCT系数在多个DCT块中从低阶重新排列成高阶。 以这种方式重新排列的DCT系数被打包成流，并且分离的报头部分被添加，形成转换的基本流。 记录转换后的基本流。 在高速播放期间，即使在宏块的中间存在错误，由于可以使用低阶DCT系数进行解码，因此可以获得高质量的图像视频数据。

公开(公告)号：US06516034B2

公开(公告)日：2003-02-04

申请号：US09350226

申请日：1999-07-08

申请人： Shin Todo ,  Haruo Togashi ,  Hideyuki Matsumoto

发明人： Shin Todo ,  Haruo Togashi ,  Hideyuki Matsumoto

IPC分类号： H04N166

CPC分类号： H04N19/00 ,  H03M7/40 ,  H04N19/124 ,  H04N19/132 ,  H04N19/149 ,  H04N19/152 ,  H04N19/162 ,  H04N19/166 ,  H04N19/174 ,  H04N19/176 ,  H04N19/177 ,  H04N19/18 ,  H04N19/40 ,  H04N19/42 ,  H04N19/46 ,  H04N19/59 ,  H04N19/61 ,  H04N19/70 ,  H04N19/89 ,  H04N19/91

摘要： A stream processing apparatus and method for replacing a macroblock having a syntax error in an encoded stream with a marcroblock having an acceptable syntax. In such apparatus and method, a plurality of coding parameters may be extracted from the source encoded stream, an error macroblock having a syntax error in the source encoded stream may be detected, the source encoded stream may be variable-length decoded so as to generate a variable-length-decoded stream, a predetermined macroblock having an acceptable syntax may be generated which conforms to a MPEG (motion picture image coding experts group) standard, and the error macroblock of the variable-length decoded stream may be replaced with the predetermined macroblock.

摘要翻译： 一种流处理装置和方法，用于用具有可接受语法的marcroblock替换具有编码流中的语法错误的宏块。 在这种装置和方法中，可以从源编码流中提取多个编码参数，可以检测到源编码流中具有语法错误的误差宏块，源编码流可以是可变长度解码的，以便产生 可以生成符合MPEG（运动图像图像编码专家组）标准的可变长度解码流，具有可接受语法的预定宏块，并且可变长度解码流的误差宏块可以用预定的 宏块。

公开(公告)号：US08155398B2

公开(公告)日：2012-04-10

申请号：US12013050

申请日：2008-01-11

申请人： Haruo Togashi

发明人： Haruo Togashi

IPC分类号： G06K9/00

CPC分类号： G06K9/527 ,  G06K9/00268 ,  G06K9/00288 ,  G06K9/6256

摘要： An image processing apparatus includes an image conversion section that receives an input of a face image to be identified, executes an image conversion on the input face image, and performs a normalization processing into an image. The image conversion section obtains a face image from a first memory storing the face image to be normalization processed, performs the normalization processing by an image conversion and stores the face image after the normalization processing into a second memory. The image processing apparatus includes a calculation section that calculates a conversion parameter for calculating a corresponding point in the first memory to each pixel position in the second memory. The conversion parameter defines one of an image contraction processing, an image rotation processing, or an image translation processing to be performed when the face image stored in the first memory is converted into the face image stored in the second memory.

摘要翻译： 图像处理装置包括图像转换部，其接收要识别的面部图像的输入，对输入的面部图像执行图像变换，并对图像进行归一化处理。 图像转换部从存储要进行归一化处理的面部图像的第一存储器获取面部图像，通过图像转换进行归一化处理，并将归一化处理后的面部图像存储到第二存储器中。 图像处理装置包括：计算部，其计算用于计算第一存储器中对应点到第二存储器中的每个像素位置的转换参数。 转换参数定义当存储在第一存储器中的面部图像被转换成存储在第二存储器中的面部图像时要执行的图像收缩处理，图像旋转处理或图像转换处理中的一个。