公开(公告)号：US20060110040A1

公开(公告)日：2006-05-25

申请号：US11327913

申请日：2006-01-09

申请人： Patrice Simard ,  John Platt ,  David Steinkraus

发明人： Patrice Simard ,  John Platt ,  David Steinkraus

IPC分类号： G06K9/62 ,  G06K9/00 ,  G06K9/18 ,  G06K9/46

CPC分类号： G06K9/4628 ,  G06K9/00422 ,  G06K2209/011

摘要： A system and method facilitating pattern recognition is provided. The invention includes a pattern recognition system having a convolutional neural network employing feature extraction layer(s) and classifier layer(s). The feature extraction layer(s) comprises convolutional layers and the classifier layer(s) comprises fully connected layers. The pattern recognition system can be trained utilizing a calculated cross entropy error. The calculated cross entropy error is utilized to update trainable parameters of the pattern recognition system.

公开(公告)号：US20070003142A1

公开(公告)日：2007-01-04

申请号：US11173243

申请日：2005-07-01

申请人： Patrice Simard ,  Maneesh Agrawala ,  David Steinkraus

发明人： Patrice Simard ,  Maneesh Agrawala ,  David Steinkraus

IPC分类号： G06K9/00

CPC分类号： G06K9/00416

摘要： Systems and methods are disclosed that facilitate normalizing and beautifying digitally generated handwriting, such as can be generated on a tablet PC or via scanning a handwritten document. A classifier can identify extrema in the digital handwriting and label such extrema according to predefined categories (e.g., bottom, baseline, midline, top, other, . . . ). Multi-linear regression, polynomial regression, etc., can be performed to align labeled extrema to respective and corresponding desired points as indicated by the labels. Additionally, displacement techniques can be applied to the regressed handwriting to optimize legibility for reading by a human viewer and/or for character recognition by a handwriting recognition application. The displacement techniques can comprise a “rubber sheet” displacement algorithm in conjunction with a “rubber rod” displacement algorithm, which can collectively preserve spatial features of the handwriting during warping thereof.

摘要翻译： 公开了促进数字生成的笔迹的归一化和美化的系统和方法，诸如可以在平板PC上生成或通过扫描手写文档。 分类器可以根据预定类别（例如，底部，基线，中线，顶部，其他等）识别数字手写中的极值并标记这样的极值。 可以执行多线性回归，多项式回归等，以将标记的极值与标签所示的相应和对应的期望点对齐。 此外，位移技术可以应用于回归的笔迹，以优化由人类观察者阅读的可读性和/或通过手写识别应用的字符识别。 位移技术可以包括“橡胶片”位移算法，结合“橡胶棒”位移算法，其可以在其翘曲期间共同保留笔迹的空间特征。

公开(公告)号：US20050125369A1

公开(公告)日：2005-06-09

申请号：US10837382

申请日：2004-04-30

申请人： Ian Buck ,  Patrice Simard ,  David Steinkraus

发明人： Ian Buck ,  Patrice Simard ,  David Steinkraus

IPC分类号： G06T1/40 ,  G06F3/14 ,  G06F15/18 ,  G06N3/00 ,  G06N3/063 ,  G06N3/08 ,  G06F17/00 ,  G06N5/02

CPC分类号： G06K9/00986 ,  G06N3/08

摘要： A system and method for processing machine learning techniques (such as neural networks) and other non-graphics applications using a graphics processing unit (GPU) to accelerate and optimize the processing. The system and method transfers an architecture that can be used for a wide variety of machine learning techniques from the CPU to the GPU. The transfer of processing to the GPU is accomplished using several novel techniques that overcome the limitations and work well within the framework of the GPU architecture. With these limitations overcome, machine learning techniques are particularly well suited for processing on the GPU because the GPU is typically much more powerful than the typical CPU. Moreover, similar to graphics processing, processing of machine learning techniques involves problems with solving non-trivial solutions and large amounts of data.

摘要翻译： 一种用于处理机器学习技术（例如神经网络）和使用图形处理单元（GPU）来加速和优化处理的其他非图形应用的系统和方法。 该系统和方法传输一种可用于从CPU到GPU的各种机器学习技术的架构。 处理到GPU的转移是通过克服这些限制并在GPU架构的框架内工作良好的几种新技术实现的。 由于克服了这些限制，机器学习技术特别适用于GPU上的处理，因为GPU通常比典型的CPU功能更强大。 此外，类似于图形处理，机器学习技术的处理涉及解决非平凡解决方案和大量数据的问题。

公开(公告)号：US20060078210A1

公开(公告)日：2006-04-13

申请号：US11287671

申请日：2005-11-28

申请人： Patrice Simard ,  Henrique Malvar ,  Dinei Florencio ,  David Steinkraus

发明人： Patrice Simard ,  Henrique Malvar ,  Dinei Florencio ,  David Steinkraus

IPC分类号： G06K9/36

CPC分类号： G06T9/004 ,  G06T9/007

摘要： Systems and methods for performing adaptive filtering are disclosed. The present invention generates probabilities that can be used in an encoder, such as an arithmetic encoder and generates those probabilities in a computationally efficient manner. Probabilities of previously encoded coefficients are employed, effectively, in generating probabilities of the coefficients without regard to directional information. Thus, a large amount of information is adaptively and efficiently used in generating the probabilities. For the coefficients, the probability is computed based at least partly on at least one probability of a previously computed probability of a neighboring coefficient. Then, the coefficients are encoded using those computed probabilities.

摘要翻译： 公开了用于执行自适应滤波的系统和方法。 本发明产生可以在诸如算术编码器的编码器中使用的概率，并以计算有效的方式生成这些概率。 先前编码的系数的概率被有效地用于在不考虑方向信息的情况下生成系数的概率。 因此，在生成概率时自适应地有效地使用大量的信息。 对于系数，概率至少部分地基于先前计算的相邻系数的概率的至少一个概率来计算。 然后，使用那些计算的概率对系数进行编码。

公开(公告)号：US20070211064A1

公开(公告)日：2007-09-13

申请号：US11748474

申请日：2007-05-14

申请人： Ian Buck ,  Patrice Simard ,  David Steinkraus

发明人： Ian Buck ,  Patrice Simard ,  David Steinkraus

IPC分类号： G06F13/14

CPC分类号： G06N99/005 ,  G06N3/08

摘要： A system and method for processing machine learning techniques (such as neural networks) and other non-graphics applications using a graphics processing unit (GPU) to accelerate and optimize the processing. The system and method transfers an architecture that can be used for a wide variety of machine learning techniques from the CPU to the GPU. The transfer of processing to the GPU is accomplished using several novel techniques that overcome the limitations and work well within the framework of the GPU architecture. With these limitations overcome, machine learning techniques are particularly well suited for processing on the GPU because the GPU is typically much more powerful than the typical CPU. Moreover, similar to graphics processing, processing of machine learning techniques involves problems with solving non-trivial solutions and large amounts of data.

摘要翻译： 一种用于处理机器学习技术（例如神经网络）和使用图形处理单元（GPU）来加速和优化处理的其他非图形应用的系统和方法。 该系统和方法传输一种可用于从CPU到GPU的各种机器学习技术的架构。 处理到GPU的转移是通过克服这些限制并在GPU架构的框架内工作良好的几种新技术实现的。 由于克服了这些限制，机器学习技术特别适用于GPU上的处理，因为GPU通常比典型的CPU功能更强大。 此外，类似于图形处理，机器学习技术的处理涉及解决非平凡解决方案和大量数据的问题。

公开(公告)号：US20070086655A1

公开(公告)日：2007-04-19

申请号：US11250819

申请日：2005-10-14

申请人： Patrice Simard ,  David Steinkraus ,  Kumar Chellapilla

发明人： Patrice Simard ,  David Steinkraus ,  Kumar Chellapilla

IPC分类号： G06K9/46 ,  G06K9/64

CPC分类号： G06K9/4628 ,  G06K2209/01

摘要： Systems and methods are described that facilitate performing feature extraction across multiple received input features to reduce computational overhead associated with feature processing related to, for instance, optical character recognition. Input feature information can be unfolded and concatenated to generate an aggregated input matrix, which can be convolved with a kernel matrix to produce output feature information for multiple output features concurrently.

摘要翻译： 描述了有助于在多个接收到的输入特征之间执行特征提取的系统和方法，以减少与例如光学字符识别相关的特征处理相关联的计算开销。 输入特征信息可以展开并连接以生成聚合输入矩阵，其可以与内核矩阵进行卷积以同时产生多个输出特征的输出特征信息。

公开(公告)号：US20050025355A1

公开(公告)日：2005-02-03

申请号：US10631511

申请日：2003-07-31

申请人： Patrice Simard ,  David Steinkraus

发明人： Patrice Simard ,  David Steinkraus

IPC分类号： G06K9/62

CPC分类号： G06K9/6256

摘要： A system that facilitates generation of data that can be employed in connection with training a classifier. The system comprises a component that receives a data set that is employed in connection with training the classifier, and an expansion component that applies elastic distortion algorithm(s) to a subset of the data set to generate additional labeled training data.

摘要翻译： 一种有助于产生可以与训练分类器结合使用的数据的系统。 该系统包括接收与训练分类器结合使用的数据集的组件以及将弹性失真算法应用于数据集的子集以产生附加标记的训练数据的扩展组件。

公开(公告)号：US20050197977A1

公开(公告)日：2005-09-08

申请号：US10877730

申请日：2004-06-25

申请人： Ian Buck ,  David Steinkraus ,  Richard Szeliski

发明人： Ian Buck ,  David Steinkraus ,  Richard Szeliski

IPC分类号： G06F15/18

CPC分类号： G06T1/20 ,  G06F17/16 ,  G06N3/063 ,  G06N99/005

摘要： A system and method for optimizing the performance of a graphics processing unit (GPU) for processing and execution of general matrix operations such that the operations are accelerated and optimized. The system and method describes the layouts of operands and results in graphics memory, as well as partitioning the processes into a sequence of passes through a macro step. Specifically, operands are placed in memory in a pattern, results are written into memory in a pattern appropriate for use as operands in a later pass, data sets are partitioned to insure that each pass fits into fixed sized memory, and the execution model incorporates generally reusable macro steps for use in multiple passes. These features enable greater efficiency and speed in processing and executing general matrix operations.

摘要翻译： 一种用于优化用于处理和执行一般矩阵运算的图形处理单元（GPU）的性能的系统和方法，使得加速和优化操作。 该系统和方法描述了图形存储器中的操作数和结果的布局，以及将进程分割成通过宏步骤的顺序。 具体来说，操作数以模式放置在存储器中，结果以适合在稍后传递中用作操作数的模式写入存储器，数据集被分区以确保每个通过符合固定大小的存储器，并且执行模型通常包含 可重复使用的宏步骤可用于多次通过。 这些特性使得在处理和执行通用矩阵运算时能够提高效率和速度。