专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明申请

US20180247193A1 NEURAL NETWORK TRAINING USING COMPRESSED INPUTS 审中-公开
公开(公告)号：US20180247193A1
公开(公告)日：2018-08-30
申请号：US15903353
申请日：2018-02-23
申请人： Xtract Technologies Inc.
发明人： Elliot Mark Holtham
IPC分类号： G06N3/08 , G06T9/00 , H04N19/60
CPC分类号： G06K9/6215 , G06F16/2365 , G06F17/5009 , G06K9/6255 , G06K9/6269 , G06K9/6857 , G06N3/0454 , G06N3/0472 , G06N3/08 , G06N3/084 , G06N5/003 , G06N7/005 , G06N20/00 , G06N20/10 , G06T9/002 , H04N19/60 , H04N19/96
摘要： Aspects relate to systems and methods for improving the operation of computer-implemented neural networks. Some aspects relate to training a neural network using a compressed representation of the inputs either through efficient discretization of the inputs, or choice of compression. This approach allows a multiscale approach where the input discretization is adaptively changed during the learning process, or the loss of the compression is changed during the training. Once a network has been trained, the approach allows for efficient predictions and classifications using compressed inputs. One approach can generate a larger more diverse training dataset based on both simulations from physical models, as well as incorporating domain expertise and other available information. One approach can automatically match the documents to the list, while still allowing a user to input information to update and correct the matching process.

2. 发明申请

US20180210774A1 IMAGE BASED FAULT STATE DETERMINATION 审中-公开
公开(公告)号：US20180210774A1
公开(公告)日：2018-07-26
申请号：US15744557
申请日：2015-09-18
申请人： HEWLETT PACKARD ENTERPRISE DEVELOPMENT lp
发明人： Erik Levon Young , Luis E. Luciani, JR.
IPC分类号： G06F11/07 , G06K9/62 , G06K9/68 , G06K9/34 , G06K9/72
CPC分类号： G06F11/079 , G06F3/14 , G06F11/0751 , G06F11/0769 , G06F11/0772 , G06F11/30 , G06F11/3006 , G06F11/32 , G06K9/344 , G06K9/6215 , G06K9/627 , G06K9/6857 , G06K9/72 , G06K2209/01 , G09G2330/12 , G09G2340/04
摘要： In some examples, a system for determining whether an operating system fault has occurred includes data storage and a processing system. The data storage may store image data indicative of a computing system display output. The processing system may access the stored image data. The processing system may determine that the computing system display output corresponds to a fault display output associated with a fault state of an operating system, which determination may include the processing system determining an extent of similarity between the accessed image data and a reference image associated with the fault display output. The processing system may generate a fault indication responsive to determining that the computing system display output corresponds to the fault display output.

3. 发明授权

US09881372B2 Method and system for vascular disease detection using recurrent neural networks 有权
公开(公告)号：US09881372B2
公开(公告)日：2018-01-30
申请号：US15680584
申请日：2017-08-18
申请人： Siemens Healthcare GmbH
发明人： Mehmet A. Gulsun , Yefeng Zheng , Puneet Sharma , Bogdan Georgescu , Dorin Comaniciu
IPC分类号： G06K9/00 , G06T7/00 , G06N3/02 , G06T7/60 , G06K9/46
CPC分类号： G06T7/0012 , G06K9/00147 , G06K9/4628 , G06K9/6271 , G06K9/6857 , G06N3/02 , G06T7/60 , G06T2200/04 , G06T2207/10081 , G06T2207/20084 , G06T2207/30101 , G06T2207/30172
摘要： A method and apparatus for vascular disease detection and characterization using a recurrent neural network (RNN) is disclosed. A plurality of 2D cross-section image patches are extracted from a 3D computed tomography angiography (CTA) image, each extracted at a respective sampling point along a vessel centerline of a vessel of interest in the 3D CTA image. Vascular abnormalities in the vessel of interest are detected and characterized by classifying each of the sampling points along the vessel centerline based on the plurality of 2D cross-section image patches using a trained RNN.

4. 发明授权

US09805443B2 Image processing method, image processing apparatus, program, storage medium, production apparatus, and method of producing assembly 有权
公开(公告)号：US09805443B2
公开(公告)日：2017-10-31
申请号：US14793431
申请日：2015-07-07
申请人： CANON KABUSHIKI KAISHA
发明人： Hiroshi Kitajima
IPC分类号： G06K9/62 , G06T1/60 , G06T1/00 , G06T1/20 , G06K9/68 , G06T7/73
CPC分类号： G06T1/60 , G06K9/6202 , G06K9/6212 , G06K9/6857 , G06T1/0007 , G06T1/20 , G06T7/74 , G06T2207/20016
摘要： A tentative local score between a point in a feature image in a template image and a point, in a target object image, at a position corresponding to the point in the feature image is calculated, and a determination is performed as to whether the tentative local score is smaller than 0. In a case where the tentative local score is greater than or equal to 0, the tentative local score is employed as a local score. In a case where the tentative local score is smaller than 0, the tentative local score is multiplied by a coefficient and the result is employed as a degree of local similarity.

5. 发明授权

US09767557B1 Method and system for vascular disease detection using recurrent neural networks 有权
公开(公告)号：US09767557B1
公开(公告)日：2017-09-19
申请号：US15429806
申请日：2017-02-10
申请人： Siemens Healthcare GmbH
发明人： Mehmet A. Gulsun , Yefeng Zheng , Puneet Sharma , Bogdan Georgescu , Dorin Comaniciu
IPC分类号： G06K9/00 , G06T7/00 , G06K9/46 , G06T7/60 , G06N3/02
CPC分类号： G06T7/0012 , G06K9/00147 , G06K9/4628 , G06K9/6271 , G06K9/6857 , G06N3/02 , G06T7/60 , G06T2200/04 , G06T2207/10081 , G06T2207/20084 , G06T2207/30101 , G06T2207/30172
摘要： A method and apparatus for vascular disease detection and characterization using a recurrent neural network (RNN) is disclosed. A plurality of 2D cross-section image patches are extracted from a 3D computed tomography angiography (CTA) image, each extracted at a respective sampling point along a vessel centerline of a vessel of interest in the 3D CTA image. Vascular abnormalities in the vessel of interest are detected and characterized by classifying each of the sampling points along the vessel centerline based on the plurality of 2D cross-section image patches using a trained RNN.

6. 发明授权

US09721334B2 Work-piece defect inspection via optical images and CT images 有权
公开(公告)号：US09721334B2
公开(公告)日：2017-08-01
申请号：US14957662
申请日：2015-12-03
申请人： International Business Machines Corporation
发明人： Boliang Chen , Min Gong , Dong Sheng Li , Junchi Yan , Chao Zhang , Wei Peng Zhang
IPC分类号： G06T7/00 , G06K9/62 , H04N5/232
CPC分类号： G06K9/6256 , G06K9/4628 , G06K9/6273 , G06K9/6857 , G06K2209/05 , G06T7/001 , G06T2207/10081 , G06T2207/10116 , G06T2207/20016 , G06T2207/30164
摘要： A computer-implemented method, computer program product, and system for detecting work-piece defects. The computer-implemented method may include: receiving a first image of a training work-piece captured using a non-destructive imaging process and a second image of the training work-piece captured using a destructive imaging process; receiving an image of a work-piece captured using a non-destructive imaging process, wherein the work-piece is substantially similar to the training work-piece, and the non-destructive imaging process used to capture the work-piece is substantially similar to the non-destructive imaging process used to capture the training work-piece; matching the image of the work-piece to the first image of the training work-piece; and enhancing the image of the work-piece using the second image of the training work-piece, in response to the image of the work-piece matching the first image of the training work-piece.

7. 发明授权

US09691127B2 Method, apparatus and computer program product for alignment of images 有权
公开(公告)号：US09691127B2
公开(公告)日：2017-06-27
申请号：US14896782
申请日：2013-06-14
申请人： Nokia Technologies Oy
发明人： Alexander Lindskog
IPC分类号： G06T3/00 , G06K9/62 , G06K9/68 , G06K9/46 , G06T3/40 , G06T7/32
CPC分类号： G06T3/0068 , G06K9/4609 , G06K9/4652 , G06K9/6204 , G06K9/6215 , G06K9/6857 , G06T3/4038 , G06T7/32 , G06T2207/10144 , G06T2207/10148 , G06T2207/10152 , G06T2207/20016 , G06T2207/20208
摘要： In an example embodiment a method, apparatus and computer program product are provided. The method includes calculating directionality values for pixels of a first image and pixels of a second image, where a directionality value for a pixel is calculated based on gradient differences between the pixel and a plurality of neighboring pixels. The method includes determining a plurality of similarity values between the first image and the second images for a plurality of alignment positions of the first image and the second image based on the directionality values for the pixels of the first image and the directionality values for corresponding pixels of the second image. The method further includes selecting an alignment position from among the plurality of alignment positions for aligning the first image and the second image based on comparison of the plurality of similarity values.

8. 发明授权

US09684967B2 Imaging segmentation using multi-scale machine learning approach 有权
公开(公告)号：US09684967B2
公开(公告)日：2017-06-20
申请号：US14921579
申请日：2015-10-23
申请人： International Business Machines Corporation
发明人： Mani Abedini , Rajib Chakravorty , Rahil Garnavi , Munawar Hayat
IPC分类号： G06K9/00 , G06T7/00 , G06K9/62 , G06K9/46 , G06K9/52
CPC分类号： G06T7/13 , G06K9/4604 , G06K9/52 , G06K9/6256 , G06K9/6262 , G06K9/6267 , G06K9/6857 , G06T7/0012 , G06T7/10 , G06T7/11 , G06T7/143 , G06T7/90 , G06T2207/20016 , G06T2207/20021 , G06T2207/20076 , G06T2207/20081 , G06T2207/20112 , G06T2207/30088 , G06T2207/30096
摘要： A robust segmentation technique based on multi-layer classification technique to identify the lesion boundary is described. The inventors have discovered a technique based on training several classifiers such that to classify each pixel as lesion versus normal Each classifier is trained on a specific range of image resolutions. Then, for a new test image, the trained classifiers are applied on the image. Then by fusing the prediction results in pixel level a probability map is generated. In the next step, a thresholding method is applied to convert the probability map to a binary mask, which determines a mole border.

9. 发明申请

US20170116744A1 IMAGING SEGMENTATION USING MULTI-SCALE MACHINE LEARNING APPROACH 有权
公开(公告)号：US20170116744A1
公开(公告)日：2017-04-27
申请号：US14921579
申请日：2015-10-23
申请人： International Business Machines Corporation
发明人： Mani ABEDINI , Rajib CHAKRAVORTY , Rahil GARNAVI , Munawar HAYAT
IPC分类号： G06T7/00 , G06K9/46 , G06K9/52 , G06K9/62
CPC分类号： G06T7/13 , G06K9/4604 , G06K9/52 , G06K9/6256 , G06K9/6262 , G06K9/6267 , G06K9/6857 , G06T7/0012 , G06T7/10 , G06T7/11 , G06T7/143 , G06T7/90 , G06T2207/20016 , G06T2207/20021 , G06T2207/20076 , G06T2207/20081 , G06T2207/20112 , G06T2207/30088 , G06T2207/30096
摘要： A robust segmentation technique based on multi-layer classification technique to identify the lesion boundary is described. The inventors have discovered a technique based on training several classifiers such that to classify each pixel as lesion versus normal Each classifier is trained on a specific range of image resolutions. Then, for a new test image, the trained classifiers are applied on the image. Then by fusing the prediction results in pixel level a probability map is generated. In the next step, a thresholding method is applied to convert the probability map to a binary mask, which determines a mole border.

10. 发明授权

US09571694B2 Image reading apparatus, method for controlling the same, and storage medium 有权
标题翻译：图像读取装置，其控制方法和存储介质
公开(公告)号：US09571694B2
公开(公告)日：2017-02-14
申请号：US14729980
申请日：2015-06-03
申请人： CANON KABUSHIKI KAISHA
发明人： Toyomi Hirao
IPC分类号： H04N1/04 , H04N1/393 , G06K9/68
CPC分类号： H04N1/0443 , G06K9/6857 , G06K2209/01 , H04N1/00822 , H04N1/3935 , H04N2201/0081 , H04N2201/0094
摘要： A resolution of an image data needs to be higher than a predetermined value for character recognition. A control method in an image reading apparatus having a reading unit for reading an image of a document to generate image data, and a storing unit for storing the image data generated by the reading unit includes receiving from a user, after the image data is stored by the storing unit, an instruction for executing character recognition processing on the image data stored by the storing unit, executing character recognition processing on the image data stored by the storing unit according to a reception of the instruction by the receiving unit, and performing control, in a case where the image data generated by the reading unit is to be stored by the storing unit, to read an image of a document by the reading unit with a resolution that enables character recognition.
摘要翻译：图像数据的分辨率需要高于用于字符识别的预定值。一种具有用于读取文档的图像以生成图像数据的读取单元的图像读取装置中的控制方法，以及用于存储由读取单元生成的图像数据的存储单元，包括在存储图像数据之后从用户接收由所述存储单元执行用于对由所述存储单元存储的图像数据执行字符识别处理的指令，对由所述存储单元存储的图像数据执行字符识别处理，根据所述接收单元的所述指令的接收，并执行控制在由存储单元存储由读取单元生成的图像数据的情况下，通过读取单元以能够进行字符识别的分辨率来读取文档的图像。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式