公开(公告)号：US06643677B2

公开(公告)日：2003-11-04

申请号：US09503080

申请日：2000-02-14

申请人： Toshiyuki Furusawa ,  Yusei Itaya ,  Masaru Ozeki

发明人： Toshiyuki Furusawa ,  Yusei Itaya ,  Masaru Ozeki

IPC分类号： G06F738

CPC分类号： G06F7/5443

摘要： A digital signal processor (DSP) of high speed and high precision is disclosed. The DSP (i.e., digital arithmetic integrated circuit) comprises: an arithmetic data storing memory (11) for storing arithmetic data and for outputting, in one instruction cycle, first and second arithmetic data strings each composed of a plurality of bits, the first arithmetic data string being composed of m-bits of at least a predetermined unit word length and the second arithmetic data string being composed of (m×n)-bits of a unit word length; two arithmetic operand storing registers (12) for storing the first and second arithmetic data strings outputted by said arithmetic data storing memory, respectively; an arithmetic logical unit for executing arithmetic operation on the basis of the two operands outputted by said arithmetic operand storing registers in one instruction cycle (13); and an arithmetic result storing register (15) for storing the arithmetic results outputted by said arithmetic logical unit.

摘要翻译： 公开了一种高速，高精度的数字信号处理器（DSP）。 DSP（即，数字运算集成电路）包括：运算数据存储存储器，用于存储运算数据，并在一个指令周期内输出由多个位组成的第一和第二运算数据串，第一运算 数据串由至少预定单位字长的m位构成，第二算术数据串由单位字长度的（m×n）位构成; 两个算术运算存储寄存器（12），用于分别存储由所述算术数据存储存储器输出的第一和第二算术数据串; 算术逻辑单元，用于在一个指令周期（13）中基于由所述算术运算存储寄存器输出的两个操作数执行算术运算; 以及用于存储由所述算术逻辑单元输出的算术结果的算术结果存储寄存器（15）。

公开(公告)号：US6070180A

公开(公告)日：2000-05-30

申请号：US658999

申请日：1996-06-04

申请人： Toshiyuki Furusawa ,  Yusei Itaya ,  Masaru Ozeki

发明人： Toshiyuki Furusawa ,  Yusei Itaya ,  Masaru Ozeki

IPC分类号： G06F7/52 ,  G06P7/38

CPC分类号： G06F7/5324 ,  G06F2207/382

摘要： A digital signal processor (DSP) of high speed and high precision is disclosed. The DSP (i.e., digital arithmetic integrated circuit) comprises: an arithmetic data storing memory (11) for storing arithmetic data and for outputting, in one instruction cycle, first and second arithmetic data strings each composed of a plurality of bits, the first arithmetic data string being composed of m-bits of at least a predetermined unit wiord length and the second arithmetic data string being composed of (m.times.n)-bits of a unit word length; two arithmetic operand storing registers (12) for storing the first and second arithmetic data strings outputted by said arithmetic data storing memory, respectively; an arithmetic logical unit for executing arithmetic operation on the basis of the two operands outputted by said arithmetic operand storing registers in one instruction cycle (13); and an arithmetic result storing register (15) for storing the arithmetic results outputted by said arithmetic logical unit.

摘要翻译： 公开了一种高速，高精度的数字信号处理器（DSP）。 DSP（即，数字运算集成电路）包括：运算数据存储存储器，用于存储运算数据，并在一个指令周期内输出由多个位组成的第一和第二运算数据串，第一运算 数据串由至少预定单位长度的m位组成，第二运算数据串由单位字长度的（m×n）位构成; 两个算术运算存储寄存器（12），用于分别存储由所述算术数据存储存储器输出的第一和第二算术数据串; 算术逻辑单元，用于在一个指令周期（13）中基于由所述算术运算存储寄存器输出的两个操作数执行算术运算; 以及用于存储由所述算术逻辑单元输出的算术结果的算术结果存储寄存器（15）。

公开(公告)号：US5802210A

公开(公告)日：1998-09-01

申请号：US488440

申请日：1995-06-07

申请人： Masanori Kurata ,  Ryusuke Murakami ,  Yusei Itaya

发明人： Masanori Kurata ,  Ryusuke Murakami ,  Yusei Itaya

IPC分类号： G06T3/40 ,  G06K9/36 ,  G06K9/46

CPC分类号： G06T3/4084

摘要： An image processing system conducts a video effect processing. In the mosaic-image-processing, a dc component of inversely quantized discrete cosine transform (DCT) coefficients is multiplied by a constant h other than zero and an ac component of the inversely quantized DCT coefficients is replaced with zero. In the case of monochromatic-image-processing, a dc component D.sub.y of a luminance block Y of the inversely quantized DCT coefficients is multiplied by a constant i other than zero and an ac component A.sub.y of the luminance block Y is multiplied by a constant I other than zero, dc components D.sub.Cb and D.sub.Cr of chrominance blocks Cb and Cr of the inversely quantized DCT coefficients are replaced with constants j and k, respectively, and ac components A.sub.Cb and A.sub.Cr of the chrominance blocks Cb and Cr are replaced with constants J and K, respectively. In the thermographic-image processing, a dc component D.sub.y of the luminance block Y is processed into dc components D.sub.Cb '=mD.sub.y +n and D.sub.Cr '=pD.sub.y +q where m, n, p and q are a constant, respectively, ac component A.sub.y of the luminance block Y is processed into ac components A.sub.Cb '=MA.sub.y +N and A.sub.Cr '=PA.sub.y +Q where M, N, P and Q are a constant, respectively, and the dc component D.sub.y and the ac component A.sub.y are replaced with constants s and S, respectively.

摘要翻译： 图像处理系统进行视频效果处理。 在马赛克图像处理中，将逆量化离散余弦变换（DCT）系数的直流分量乘以除零以外的常数h，并将逆量化的DCT系数的ac分量替换为零。 在单色图像处理的情况下，将逆量化DCT系数的亮度块Y的直流分量Dy乘以除零之外的常数i，将亮度块Y的交流分量Ay乘以常数I 除零之外，反向量化DCT系数的色度块Cb和Cr的直流分量DCb和DCr分别被常数j和k代替，色度块Cb和Cr的ac分量ACb和ACr被代入常数J和 K。 在热成像图像处理中，亮度块Y的直流分量Dy被处理为直流分量DCb'= mDy + n，DCr'= pDy + q，其中m，n，p和q分别为常数，ac分量 亮度块Y的Ay分别被处理为ac分量ACb'= MAy + N和ACr'= PAy + Q，其中M，N，P和Q分别是常数，并且替换直流分量Dy和ac分量Ay 分别具有常数s和S。