公开(公告)号：EP1615168A1

公开(公告)日：2006-01-11

申请号：EP04425503.2

申请日：2004-07-09

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe ,  Bruna, Arcagenlo Ranieri ,  Sanguedolce, Giuseppe

IPC: G06T3/40

CPC classification number: G06T3/4084

Abstract: A method of color interpolation of pixels of an image detected by a digital color sensor comprises the steps of interpolating separately the pixels belonging to each of a plurality of subdivision channels of the pixels of the detected image respectively for primary red-green-blue colors (RGB) or for base hues for generating weighted values relating to either said primary colors or to said base hues for each pixel of the image to be output.
This method generates images with enhanced details because it comprises the steps of defining wavelet coefficients of the wavelet sub-band LL depending on the values of the pixel of each channel; and estimating for each channel wavelet coefficients of the other wavelet sub-bands HL, LH and HH of the image to be output in function of said wavelet coefficients of the wavelet sub-band LL of the image to be output.

Abstract translation: 由数字颜色传感器检测的图像的像素的颜色插值的方法包括以下步骤：分别对属于检测图像的像素的多个细分通道中的每一个的像素分别插入用于初级红绿蓝色（ RGB）或用于生成与要输出的图像的每个像素相关的所述原色或所述基色调的加权值的基色调。 该方法产生具有增强细节的图像，因为它包括根据每个通道的像素的值来定义小波子带LL的小波系数的步骤; 并且根据要输出的图像的小波子带LL的所述小波系数的函数，估计要输出的图像的其他小波子带HL，LH和HH的每个信道小波系数。

公开(公告)号：EP1622393A1

公开(公告)日：2006-02-01

申请号：EP04425587.5

申请日：2004-07-30

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe ,  Messina, Giuseppe ,  Bruna, Arcangelo Ranieri ,  Guarnera, Mirko

IPC: H04N9/04 ,  G06T3/40

CPC classification number: G06T3/4015 ,  H04N9/045 ,  H04N2209/046

Abstract: A fast method of color interpolation of pixels of an image acquired by a color filtered digital sensor uses a very simple cost function that nevertheless produce interpolated images of good quality. The cost function is computationally simpler because it does not require the calculation of powers and square roots.
An efficient triangulation algorithm that may be executed in far less time than prior art triangulation processes, while practically ensuring the same performances. The peculiarity of this triangulation algorithm consists in that on average it requires only two iteration steps, while the most accurate prior triangulation algorithm is completed only after four iteration steps.
Optionally, the interpolation process may be followed by an anti-aliasing processing that effectively removes color artifacts that may be created during the interpolation process.

Abstract translation: 由彩色滤波数字传感器获取的图像的像素的彩色插值的快速方法使用非常简单的成本函数，其仍然产生良好质量的内插图像。 成本函数在计算上更简单，因为它不需要功率和平方根的计算。 一种有效的三角测量算法，可以在比现有技术的三角测量过程更少的时间内执行，同时实际确保相同的性能。 这种三角测量算法的特点在于，平均来说，它只需要两个迭代步骤，而最精确的先验三角测量算法仅在四个迭代步骤之后完成。 可选地，内插处理之后可以是抗锯齿处理，其有效地消除可能在插值处理期间产生的颜色伪像。

公开(公告)号：EP1717753A2

公开(公告)日：2006-11-02

申请号：EP06112972.2

申请日：2006-04-24

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe ,  Capra, Alessandro ,  Pappalardo, Francesco

IPC: G06T1/60

CPC classification number: G06T1/60

Abstract: An image generating pipeline (IGP) includes a digital signal processor for implementing processing blocks connected in cascade for processing an input image that includes an array of raw pixel values to generate a color image that includes an array of reconstructed pixel values. A memory is coupled to the digital signal processor for storing the raw pixel values and the array of reconstructed pixel values. The digital signal processor includes a data cache, and the raw pixel values of the input image are processed through the processing blocks in sub-arrays having fractional dimensions of the pixel dimensions of the whole image array. The sub-arrays include an input sub-array of pixel values being loaded from the memory for defining a working window. The sub-arrays of raw pixel values have a row-wise dimension of at least a fraction of a full row of the input image, and a column-wise dimension equal to or larger than a column-wise filtering action of a respective processing block to which the input sub-array is input. The digital signal processor outputs at least one fraction of full rows of completely reconstructed pixel values of the input image for storing in the memory.

Abstract translation: 图像生成流水线（IGP）包括数字信号处理器，用于实现级联连接的处理块，用于处理包括原始像素值阵列的输入图像，以生成包括重构像素值阵列的彩色图像。 存储器耦合到数字信号处理器，用于存储原始像素值和重构像素值的阵列。 数字信号处理器包括数据高速缓存，并且通过具有整个图像阵列的像素尺寸的分数维的子阵列中的处理块来处理输入图像的原始像素值。 子阵列包括从存储器加载的用于定义工作窗口的像素值的输入子阵列。 原始像素值的子阵列具有输入图像的完整行的至少一部分的行方向尺寸，以及等于或大于相应处理块的逐列滤波动作的列方向尺寸 输入子数组的输入。 数字信号处理器输出用于存储在存储器中的输入图像的完整重建的像素值的至少一部分的全行。

公开(公告)号：EP1643441A1

公开(公告)日：2006-04-05

申请号：EP05108265.9

申请日：2005-09-08

Applicant: STMicroelectronics S.r.l.

Inventor： Messina, Giuseppe ,  Guarnera, Mirko Ignazio ,  Tomaselli, Valeria ,  Bruna, Arcangelo Ranieri ,  Spampinato, Giuseppe ,  Castorina, Alfio

IPC: G06T3/40 ,  G06T5/00 ,  G06T5/20

CPC classification number: G06T3/4015

Abstract: Subdivision per basic color channels of grey level data generated by a color sensor is no longer required according to a novel color interpolation method of an image acquired by a digital color sensor generating grey levels for each image pixel in function of the filter applied to the sensor, by interpolating the values of missing colors of each image pixel for generating triplets or pairs of values of primary colors (RGB) or complementary hues for each image pixel. The method comprises the following steps:

i) calculating spatial variation gradients of primary colors or complementary hues for each image pixel and storing the information of directional variation of primary color or complementary hue in look-up tables pertaining to each pixel;
ii) interpolating color values of each image pixel considering said directional variation information of the respective values of primary colors or complementary hues stored in the respective look-up tables of the pixel for generating said multiple distinct values for each image pixel.

Abstract translation: 每通过一个颜色传感器产生的灰度级数据的基本颜色通道分区不再需要雅丁施加到传感器由数字彩色传感器生成在过滤器的功能对于每个图像像素的灰度级获取的图像的新颖的色彩插值方法 ，通过内插每个图像像素的颜色缺失的值，用于产生原色（RGB）或互补的色调值的三元组或对每个图像像素。 该方法包括以下步骤：i）计算的基色的或互补的色调空间变化梯度为每个图像像素和存储在属于每个像素查找表原色或互补色调的定向变化的信息; ⅱ）内插考虑的原色或存储在像素的respectivement查找表为每个图像像素产生所述多个不同值的色调互补的respectivement值的所述定向变化信息的每个图像像素的色彩值。

公开(公告)号：EP1484927A2

公开(公告)日：2004-12-08

申请号：EP04011733.5

申请日：2004-05-18

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe

IPC: H04N7/64 ,  H04N7/50

CPC classification number: H04N19/61 ,  H04N19/68 ,  H04N19/69 ,  H04N19/895

Abstract: Method (Img Dec ) for decoding-decompressing a compressed-encoded digital data sequence (Bit Stream ) relating to at least one initial digital image (Img n ) comprising the steps of:

receiving (Rx_Parse) the digital data sequence (Bit Stream ) comprising compressed-encoded data groups separated from one another by at least one restart marker and each one including a respective plurality of encoded data structures,
calculating (RM_Detect) a representative value of the number of encoded data structures comprised between a first restart marker (RM 1n ) and a subsequent second restart marker signalling, respectively, the start of a first data group (VP 1n ) to be decoded and the start of a second data group (VP 2n ),
extracting (MB_Extract) from the first data group the encoded data structures;
   detecting (MB_Check) the presence of at least one error, if the number of the encoded data structures extracted is different from the calculated value.

Abstract translation: 方法（ImgDec），用于解码解压缩与至少一个初始的数字图像（IMGN）包括以下步骤压缩编码的数字数据序列（比特流）：接收（Rx_Parse）的数字数据序列（比特流），其包括压缩编码的数据 基团通过至少一个重新开始标记和每一个包括被编码的数据结构的一个respectivement多元性，计算（RM_Detect）第一重新开始标志（RM1n）和随后的第二之间包含的编码数据的结构的数量的代表值彼此分开 重新开始标记信号，分别，第一数据组（VP1n）的开始要被解码并从第一dataGroup时提取（MB_Extract）的编码数据结构的第二数据组（VP2N）的开始; 检测（MB_Check）至少一个错误的情况下，如果所提取的编码后的数据结构的数目是从计算出的值不同。

公开(公告)号：EP1622393B1

公开(公告)日：2010-03-31

申请号：EP04425587.5

申请日：2004-07-30

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe ,  Messina, Giuseppe ,  Bruna, Arcangelo Ranieri ,  Guarnera, Mirko

IPC: H04N9/04 ,  G06T3/40

CPC classification number: G06T3/4015 ,  H04N9/045 ,  H04N2209/046

公开(公告)号：EP1605400A1

公开(公告)日：2005-12-14

申请号：EP04425433.2

申请日：2004-06-11

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe ,  Pappalardo, Francesco ,  Fodera, Paolo Antonino ,  Virlinzi, Francesco ,  Capra, Alessandro

IPC: G06T1/60 ,  G06T3/40

CPC classification number: G06T3/4015 ,  G06T1/20

Abstract: A color image pixel data processing pipeline for performing, among corrective and image enhancement steps, at least an interpolation on color data to generate triplets located at distinct pixel locations, including among a plurality of defect correction and image enhancement blocks at least a first color interpolation block, generating RGB information for each pixel of the input image pixel pattern, a second color interpolation block receiving the RGB pattern pixels from said first color interpolation block and rendering enhanced RGB pattern pixels, and a plurality of dedicated line memories and delay circuits associated to and cooperating with said interpolation and correction blocks for permitting real-time processing of pixel data, further includes input image pixel pattern data read/write buffers first and second, of identical capacity suitable to store a subset or pixel block (m*n) of the image data, for translating the scanning mode of pixel data being fed to the input line memory and delay circuits associated to al least said first color interpolation block of said pipeline, from linewise to columnwise, for each subset of data stored in said first and second buffers.
The switching from linewise writing to columnwise reading of the data temporarily stored in input buffer means, reduces overall memory requisite without sensibly increasing computational overhead.

公开(公告)号：EP1484927A3

公开(公告)日：2004-12-15

申请号：EP04011733.5

申请日：2004-05-18

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe

IPC: H04N7/64 ,  H04N7/50

CPC classification number: H04N19/61 ,  H04N19/68 ,  H04N19/69 ,  H04N19/895

Abstract: Method (Img Dec ) for decoding-decompressing a compressed-encoded digital data sequence (Bit Stream ) relating to at least one initial digital image (Img n ) comprising the steps of: receiving (Rx_Parse) the digital data sequence (BitStream) comprising compressed-encoded data groups separated from one another by at least one restart marker and each one including a respective plurality of encoded data structures, calculating (RM_Detect) a representative value of the number of encoded data structures comprised between a first restart marker (RM1n) and a subsequent second restart marker signalling, respectively, the start of a first data group (VP1n) to be decoded and the start of a second data group (VP2n), extracting (MB_Extract) from the first data group the encoded data structures;    detecting (MB_Check) the presence of at least one error, if the number of the encoded data structures extracted is different from the calculated value.

公开(公告)号：EP1482449A2

公开(公告)日：2004-12-01

申请号：EP04425345.8

申请日：2004-05-14

Applicant: STMicroelectronics S.r.l.

Inventor： Spampinato, Giuseppe

IPC: G06T9/00

CPC classification number: G06T9/005 ,  H04N19/44 ,  H04N19/65 ,  H04N19/895

Abstract: Method (Img Dec ) for decoding-decompressing a compressed-encoded digital data sequence (Bit Stream ) relating to at least one compressed-encoded digital image (Img n ) and for providing at least one respective decoded-decompressed digital image.

Abstract translation: 该方法涉及处理数字图像序列，用于获得一组具有数字值CORRESPONDING的到解量化系数在一个两维空间频率域变换解量化块。 具有数字值的预设范围之外的数字值的块被选择和丢弃。 阿具有respectivement像素块替换信息以解码-解压缩的图像提供。 因此独立权利要求中包括了以下内容：（a）从通信终端传递环的初始数字图像的方法; （B）一个多媒体通信装置。