公开(公告)号：US5737451A

公开(公告)日：1998-04-07

申请号：US630553

申请日：1996-04-10

申请人： Bhavan R. Gandhi ,  Majid Rabbani ,  Rodney L. Miller

发明人： Bhavan R. Gandhi ,  Majid Rabbani ,  Rodney L. Miller

IPC分类号： G06T9/00 ,  H04N7/26 ,  H04N7/30 ,  G06T5/00 ,  H04N1/409 ,  H04N1/415

CPC分类号： H04N19/86 ,  H04N19/117 ,  H04N19/14 ,  H04N19/149 ,  H04N19/154 ,  H04N19/18 ,  H04N19/44 ,  H04N19/48 ,  H04N19/527 ,  H04N19/60 ,  H04N19/146

摘要： An improved method for removing blocking artifacts in regions of slowly varying intensity in an electronic image decoded from a transform coded representation of an image, includes the steps of: receiving signals indicative of DC and AC coefficient values of transformed blocks of image data; modifying the DC value in each of the blocks based on the level of activity within the block and the DC values in neighboring blocks; performing AC prediction in each block using the modified DC values; replacing the AC values in each block with the AC predicted values based on the level of activity within the block and the modified DC values in neighboring blocks; and reconstructing the image using the modified DC and AC coefficient values.

摘要翻译： 一种用于去除从图像的变换编码表示解码的电子图像中的缓慢变化强度的区域中的块伪影的改进方法，包括以下步骤：接收指示图像数据的变换块的DC和AC系数值的信号; 基于块内的活动级别和相邻块中的DC值来修改每个块中的DC值; 使用修改的DC值在每个块中执行AC预测; 基于块内的活动水平和相邻块中的修改的DC值，用AC预测值替换每个块中的AC值; 并使用经修改的DC和AC系数值重建图像。

公开(公告)号：US5596602A

公开(公告)日：1997-01-21

申请号：US409571

申请日：1995-03-24

申请人： Douglas W. Couwenhoven ,  Bhavan R. Gandhi ,  Craig M. Smith

发明人： Douglas W. Couwenhoven ,  Bhavan R. Gandhi ,  Craig M. Smith

IPC分类号： H03M3/04 ,  H03M7/36 ,  H04N7/26 ,  H04N7/32 ,  H04N7/50 ,  H04B1/66 ,  H04B14/06

CPC分类号： H04N19/184 ,  H04N19/115 ,  H04N19/126 ,  H04N19/147 ,  H04N19/152 ,  H04N19/176 ,  H04N19/61 ,  H04N19/146 ,  H04N19/50

摘要： A data compression system can operate in a number of different configurations, and wherein the bit rate produced by a given configuration can be controlled over some finite range by a control signal from a rate controller for smoothly transitioning between the configurations so that discontinuous jumps in bit rate and distortion are minimized is provided by determining thresholds on the feedback control signals, the thresholds being used to determine when to switch in or out of each configuration; the thresholds furthermore being determined from the intersection points of the rate distortion curves for the available configurations.

摘要翻译： 数据压缩系统可以以多种不同的配置操作，并且其中由给定配置产生的比特率可以通过来自速率控制器的控制信号在一些有限范围内被控制，以在配置之间平滑地转换，使得不连续的跳跃位 通过确定反馈控制信号的阈值来提供速率和失真最小化，所述阈值用于确定何时切换或切换每个配置; 此外，还从可用配置的速率失真曲线的交点确定阈值。

公开(公告)号：US20080077606A1

公开(公告)日：2008-03-27

申请号：US11535235

申请日：2006-09-26

申请人： Jianjun Fang ,  Bhavan R. Gandhi ,  Faisal Ishtiaq ,  Alfonso Martinez Smith ,  Wei Wang

发明人： Jianjun Fang ,  Bhavan R. Gandhi ,  Faisal Ishtiaq ,  Alfonso Martinez Smith ,  Wei Wang

IPC分类号： G06F7/00

CPC分类号： H03M7/30 ,  G06F17/2252

摘要： Both an XML schema and XML instance data as correspond to an XML document are provided (301). The XML schema is processed (302) apart from the XML instance data to provide resultant compressed XML schema data while the XML instance data is processed (303) to provide a corresponding XML instance table. The latter is compressed (304) to yield a resultant compressed XML instance table. Following receipt of such items, the compressed XML instance table is decompressed (403) to provide a resultant XML instance table with the latter being used (404), along with the XML schema, to facilitate a corresponding XML document process.

摘要翻译： 提供了与XML文档对应的XML模式和XML实例数据（301）。 XML模式在XML实例数据之外处理（302），以在处理XML实例数据（303）时提供最终的压缩XML模式数据，以提供相应的XML实例表。 后者被压缩（304）以产生最终压缩的XML实例表。 在接收到这样的项目之后，压缩的XML实例表被解压缩（403）以提供结果XML实例表，后者被使用（404）以及XML模式，以便于相应的XML文档处理。

公开(公告)号：US5466918A

公开(公告)日：1995-11-14

申请号：US145051

申请日：1993-10-29

申请人： Lawrence A. Ray ,  Richard N. Ellson ,  Bhavan R. Gandhi

发明人： Lawrence A. Ray ,  Richard N. Ellson ,  Bhavan R. Gandhi

IPC分类号： G06K19/06 ,  G06K1/12 ,  G06T9/00 ,  G07C9/00 ,  G07F7/10 ,  G11B20/10 ,  H03M7/30 ,  H04N5/92 ,  H04N7/26 ,  G06K19/12

CPC分类号： G06T9/008 ,  G07C9/00079 ,  G07F7/10

摘要： A transaction card is described in which data representing an identifiable image can be stored in a limited amount of storage space. The image to be stored is converted to an associated matrix of pixel values. The matrix of pixel values is partitioned into a plurality of ordered image portions corresponding to a like partitioning of the image. The partitioning is generally related to feature locations in the image. Each image portion is compared with a reference set of image portions represented by pixel groups. Associated with each pixel group is a signal group. For each image portion, a signal group is chosen for the associated reference pixel group that most closely matches with the pixel image portion. The signal groups representing the matrix of pixel values or the image to be stored are stored as bit patterns on the storage space of a transaction card. The bit patterns may then be read and converted to the signal groups. From the signal groups, a reference pixel group is identified from within a reference set and the reference pixel group is positioned with respect to the image at the same location as the image portion. Using this technique, a recognizable image can be constructed from image information compatible with the ISO-7811/2 standards.

摘要翻译： 描述了一种交易卡，其中表示可识别图像的数据可以存储在有限量的存储空间中。 要存储的图像被转换为​​相关联的像素值矩阵。 将像素值的矩阵划分为对应于图像的相似划分的多个有序图像部分。 分区通常与图像中的特征位置有关。 每个图像部分与由像素组表示的图像部分的参考集进行比较。 与每个像素组相关联的是信号组。 对于每个图像部分，为与像素图像部分最匹配的相关联的参考像素组选择信号组。 表示像素值的矩阵的信号组或要存储的图像作为位模式存储在交易卡的存储空间上。 然后可以将位模式读取并转换成信号组。 从信号组中，从参考集合内识别参考像素组，并且参考像素组相对于与图像部分相同位置的图像定位。 使用这种技术，可以从符合ISO-7811/2标准的图像信息构建可识别的图像。

公开(公告)号：US20110179001A1

公开(公告)日：2011-07-21

申请号：US13011831

申请日：2011-01-21

申请人： Brett L. Lindsley ,  Bhavan R. Gandhi

发明人： Brett L. Lindsley ,  Bhavan R. Gandhi

IPC分类号： G06F17/30

CPC分类号： G06F16/438 ,  G06F16/435

摘要： A method 300 for retrieving media content (120). The method can include receiving metadata (118) associated with a first media file, receiving the first media file, processing at least the metadata to identify at least a second media file (124) likely to be of interest to a user, and automatically downloading the second media file. Processing the metadata can include identifying at least one parameter contained in the metadata and/or the first media file that correlates to the second media file. Metadata (122) also associated with the second media file also can be downloaded.

摘要翻译： 一种用于检索媒体内容（120）的方法300。 该方法可以包括接收与第一媒体文件相关联的元数据（118），接收第一媒体文件，至少处理元数据以识别用户可能感兴趣的至少第二媒体文件（124），并自动下载 第二个媒体文件。 处理元数据可以包括识别包含在与第二媒体文件相关的元数据和/或第一媒体文件中的至少一个参数。 还可以下载与第二媒体文件相关联的元数据（122）。

公开(公告)号：US20080059535A1

公开(公告)日：2008-03-06

申请号：US11467974

申请日：2006-08-29

申请人： Brett L. Lindsley ,  Bhavan R. Gandhi

发明人： Brett L. Lindsley ,  Bhavan R. Gandhi

IPC分类号： G06F17/30

CPC分类号： G06F16/438 ,  G06F16/435

摘要： A method 300 for retrieving media content (120). The method can include receiving metadata (118) associated with a first media file, receiving the first media file, processing at least the metadata to identify at least a second media file (124) likely to be of interest to a user, and automatically downloading the second media file. Processing the metadata can include identifying at least one parameter contained in the metadata and/or the first media file that correlates to the second media file. Metadata (122) also associated with the second media file also can be downloaded.

摘要翻译： 一种用于检索媒体内容（120）的方法300。 该方法可以包括接收与第一媒体文件相关联的元数据（118），接收第一媒体文件，至少处理元数据以识别用户可能感兴趣的至少第二媒体文件（124），并自动下载 第二个媒体文件。 处理元数据可以包括识别包含在与第二媒体文件相关的元数据和/或第一媒体文件中的至少一个参数。 还可以下载与第二媒体文件相关联的元数据（122）。

公开(公告)号：US07697752B2

公开(公告)日：2010-04-13

申请号：US11312219

申请日：2005-12-20

申请人： Dongge Li ,  Bhavan R. Gandhi ,  Zhu Li

发明人： Dongge Li ,  Bhavan R. Gandhi ,  Zhu Li

IPC分类号： G06K9/62

CPC分类号： G06K9/00248

摘要： An object detection algorithm that generates a two-layer Gaussian Mixture Model (GMM) during a training session, and subsequent to the training session, uses the two-layer GMM to perform face detection. No labeling of local features is needed. The only input that is provided by a user is the setting of a few global parameters for the image being captured during the training session, such as, for example, the person's facial pose.

摘要翻译： 在训练阶段和训练阶段之后生成双层高斯混合模型（GMM）的对象检测算法使用双层GMM进行人脸检测。 不需要标记局部特征。 由用户提供的唯一输入是在训练会话期间为例如人的面部姿势捕获的图像的几个全局参数的设置。

公开(公告)号：US5708511A

公开(公告)日：1998-01-13

申请号：US734551

申请日：1996-10-17

申请人： Bhavan R. Gandhi ,  Craig Michael Smith ,  James R. Sullivan ,  Douglas W. Couwenhoven ,  Gregory Rombola

发明人： Bhavan R. Gandhi ,  Craig Michael Smith ,  James R. Sullivan ,  Douglas W. Couwenhoven ,  Gregory Rombola

IPC分类号： H03M7/38 ,  H04N1/41 ,  H04N1/417 ,  H04N7/26 ,  H04N7/32 ,  H04N7/34 ,  H04N1/415

CPC分类号： H04N19/12 ,  H04N19/50 ,  H04N19/593 ,  H04N19/30

摘要： A block adaptive differential pulse code modulation (DPCM) system includes both a lossless DPCM processor responsive to blocks of pixel values for producing encoder command signals and a lossy DPCM compressor responsive to blocks of pixel values for producing encoder command signals. An encoder receives the encoder command signals from the lossless DPCM processor and lossy DPCM compressor and produces a compressed encoded bit stream. A switch responsive to a compression configuration signal and to the encoder command signals from the lossy compressor selectively passes the encoder command signals from the lossless processor or the lossy compressor to the encoder.

摘要翻译： 块自适应差分脉码调制（DPCM）系统包括响应于用于产生编码器命令信号的像素值块的无损DPCM处理器和响应于用于产生编码器命令信号的像素值块的有损DPCM压缩器。 编码器从无损DPCM处理器和有损DPCM压缩器接收编码器命令信号，并产生压缩的编码比特流。 响应于压缩配置信号和来自有损压缩器的编码器命令信号的开关选择性地将编码器命令信号从无损处理器或有损压缩器传递到编码器。

公开(公告)号：US5789726A

公开(公告)日：1998-08-04

申请号：US758120

申请日：1996-11-25

申请人： Lawrence A. Ray ,  Bhavan R. Gandhi ,  Gene E. Pios

发明人： Lawrence A. Ray ,  Bhavan R. Gandhi ,  Gene E. Pios

IPC分类号： G06K7/14 ,  G06T9/00 ,  G07C9/00 ,  G06K5/00

CPC分类号： G07C9/00079 ,  G06K7/14 ,  G06T9/008

摘要： A method of encoding and decoding a transaction card with a block of image data digitally representing the features of a digital portrait of an authorized transaction card user is disclosed along with apparatus for performing the method. The encoding method partitions the digital portrait of the transaction card user into feature blocks. Each feature block is compared against a library of like feature blocks to determine the best match with a pair of library feature blocks. The library's feature blocks are derived from a plurality of portraits taken from the general population. Each library feature block is represented (addressed for access) by a code-vector in a codebook. The two code-vectors corresponding to the two closest matching features blocks from the library of like feature blocks are recorded onto a transaction card along with an interstitial pointer which indicates a weight to be allocated to each of the two code-vectors to form a better match for at least one of the library feature blocks to the user's feature blocks. All of the code-vectors and the interstitial pointers corresponding to the user's feature blocks are recorded as a sequence of sub-data blocks within an image data block on the transaction card. A method and associated apparatus for reading out the image data block and recreating a representation of the portrait of the transaction card user at a financial transaction site for identification purposes are also disclosed.

摘要翻译： 公开了一种用数字表示授权交易卡用户的数字肖像的特征的图像数据块对交易卡进行编码和解码的方法以及用于执行该方法的装置。 编码方法将交易卡用户的数字肖像划分为特征块。 将每个特征块与相似特征块的库进行比较，以确定与一对库特征块的最佳匹配。 图书馆的功能块源于从一般人群拍摄的多幅肖像。 每个库特征块由码本中的代码向量表示（寻址以访问）。 与来自相似特征块的库的两个最接近的匹配特征块相对应的两个代码矢量与间隔指针一起被记录在事务卡上，间隔指针指示要分配给两个代码矢量中的每一个的权重以形成更好的 将至少一个库功能块匹配到用户的功能块。 对应于用户特征块的所有码矢量和间质指针被记录为交易卡上的图像数据块内的子数据块序列。 还公开了一种用于读出图像数据块并且为了识别目的在金融交易站点重新创建交易卡用户的肖像的表示的方法和相关联的装置。

公开(公告)号：US5583656A

公开(公告)日：1996-12-10

申请号：US999631

申请日：1992-12-31

申请人： Bhavan R. Gandhi ,  James R. Sullivan

发明人： Bhavan R. Gandhi ,  James R. Sullivan

IPC分类号： H03M7/30 ,  G06T1/20 ,  G06T5/00 ,  G06T9/00 ,  H04N1/46 ,  H04N1/60 ,  G02B23/08 ,  G02B23/00 ,  G03B21/56

CPC分类号： G06T9/00

摘要： Each member of a set of transform representative data structures (such as look-up tables (LUTs), each representing an N TO M-dimensional transform), is stored in compressed form in a storage device included in, or which is otherwise accessible to, an image processing device or system. Each stored transform representative data structure may then, according to one embodiment of the invention, be selectively attached to image data being processed by the image processing device or system. According to a further aspect of the invention, image data, having compressed transform representative data attached thereto, may be processed utilizing methods and apparatus operative to detach, decompress and apply the transform represented by the decompressed transform representative data, to a decompressed version of the image data. The result of the aforementioned process is for image data to be transformed according to a transform specified by information previously attached to the image data.

摘要翻译： 一组变换代表数据结构（诸如查找表（LUT），每个代表N TO M维度变换）的每个成员以压缩形式存储在包括在其中的存储设备中，或者以其他方式可访问 ，图像处理装置或系统。 然后，根据本发明的一个实施例，每个存储的变换代表数据结构可以选择性地附加到由图像处理设备或系统正在处理的图像数据。 根据本发明的另一方面，可以利用附加到其上的压缩变换代表数据的图像数据，使用可解压缩和解压缩并将由解压缩变换代表数据表示的变换应用于解压缩版本的方法和装置 图像数据。 上述处理的结果是根据由先前附加到图像数据的信息指定的变换来变换图像数据。