-
公开(公告)号：US10997500B1
公开(公告)日：2021-05-04
申请号：US15603037
申请日：2017-05-23
Applicant: Amazon Technologies, Inc.
Inventor： FNU Vishnu Narayanan , Oleg Rybakov , Siddharth Singh
Abstract: The present disclosure is directed to generating neural network (NN) output using input data representing various types of events, such as input representing a certain type of event and also an engagement metric that may be representative of a property of the event or representative of a related but different type of event. For example, the output values generated using the NN may be associated with the likelihood that certain future events will occur, given the occurrence of certain past or current events. The output can then be modified (e.g., re-ranked, adjusted, etc.) based on the occurrence of certain other past or current events.
-
公开(公告)号：US09898444B1
公开(公告)日：2018-02-20
申请号：US15074263
申请日：2016-03-18
Applicant: Amazon Technologies, Inc.
Inventor： Oleg Rybakov , Andre Young Moeller , Ram Prasad Venkatesan
CPC classification number: G06K9/6276 , G06K9/344 , G06K9/6202 , G06T7/0026 , G06T7/003 , G06T7/0044 , G06T7/0081 , G06T7/408 , G06T7/60 , G06T2207/30176 , H04L67/02 , H04L67/025 , H04L67/10
Abstract: Disclosed are various embodiments for comparing images of network pages using computer vision to identify changes that have occurred between two versions of a network page. A first plurality of segments in a first image representing a first version of a network page are identified and a second plurality of segments in a second image representing a second version of the network page are identified. It is then determined whether each segment in the first plurality of segments matches a respective segment in the second plurality of segments or vice versa.
-
公开(公告)号：US10896459B1
公开(公告)日：2021-01-19
申请号：US16842432
申请日：2020-04-07
Applicant: Amazon Technologies, Inc.
Inventor： Rejith George Joseph , Oleg Rybakov
Abstract: Some aspects of the present disclosure relate to generating and training a neural network by separating historical item interaction data into both inputs and outputs. This may be done, for example, based on date. For example, a neural network machine learning technique may be used to generate a prediction model using a set of inputs that includes both a number of items purchased by a number of users before a certain date as well as some or all attributes of those items, and a set of outputs that includes the items purchased after that date. The items purchased before that date and the associated attributes can be subjected to a time-decay function.
-
公开(公告)号：US08965117B1
公开(公告)日：2015-02-24
申请号：US14109204
申请日：2013-12-17
Applicant: Amazon Technologies, Inc.
Inventor： Oleg Rybakov , Christopher John Lish , Chang Yuan , Junxiong Jia , Rakesh Madhavan Nambiar , Matias Omar Gregorio Benitez
CPC classification number: G06K9/18 , G06K9/00456 , G06K9/183 , G06K9/4604 , G06K9/6227 , G06K9/6267 , G06T5/002
Abstract: Embodiments of the subject technology provide methods and systems of image pre-processing for improving the accuracy of optical character recognition (OCR) and reducing the power consumption on a given computing device (e.g., mobile computing device). The subject technology, in some examples, classifies an image received from a camera of a mobile computing device into one or more classes: 1) normal background, 2) textured background, 3) image with text, 4) image with barcode, 5) image with QR code, and/or 6) image with clutter or “garbage.” Based on the classes associated with the image, the subject technology may forgo certain image processing operations, when the image is not associated with a particular class, in order to save resources (e.g., CPU cycles, battery power, memory usage, etc.) on the mobile computing device.
Abstract translation: 主题技术的实施例提供了用于提高光学字符识别（OCR）的精度并降低给定计算设备（例如，移动计算设备）的功耗的图像预处理的方法和系统。 在一些示例中，主题技术将从移动计算设备的相机接收的图像分类为一个或多个类别：1）正常背景，2）纹理背景，3）具有文本的图像，4）具有条形码的图像，5） 具有QR码的图像和/或6）具有杂波或“垃圾”的图像。基于与图像相关联的类别，当图像与特定类别不相关时，主题技术可以放弃某些图像处理操作 以在移动计算设备上节省资源（例如，CPU周期，电池电量，存储器使用等）。
-
公开(公告)号：US10970629B1
公开(公告)日：2021-04-06
申请号：US15442453
申请日：2017-02-24
Applicant: Amazon Technologies, Inc.
Inventor： Leo Parker Dirac , Oleg Rybakov , Vijai Mohan
Abstract: The present disclosure is directed to reducing model size of a machine learning model with encoding. The input to a machine learning model may be encoded using a probabilistic data structure with a plurality of mapping functions into a lower dimensional space. Encoding the input to the machine learning model results in a compact machine learning model with a reduced model size. The compact machine learning model can output an encoded representation of a higher-dimensional space. Use of such a machine learning model can include decoding the output of the machine learning model into the higher dimensional space of the non-encoded input.
-
Patent Agency Ranking
公开(公告)号：US10650432B1
公开(公告)日：2020-05-12
申请号：US15362585
申请日：2016-11-28
Applicant: Amazon Technologies, Inc.
Inventor： Rejith George Joseph , Oleg Rybakov
Abstract: Some aspects of the present disclosure relate to generating and training a neural network by separating historical item interaction data into both inputs and outputs. This may be done, for example, based on date. For example, a neural network machine learning technique may be used to generate a prediction model using a set of inputs that includes both a number of items purchased by a number of users before a certain date as well as some or all attributes of those items, and a set of outputs that includes the items purchased after that date. The items purchased before that date and the associated attributes can be subjected to a time-decay function.
-
公开(公告)号：US09881226B1
公开(公告)日：2018-01-30
申请号：US14864605
申请日：2015-09-24
Applicant: Amazon Technologies, Inc.
Inventor： Oleg Rybakov , Matias Omar Gregorio Benitez , Leo Parker Dirac , Rejith George Joseph , Vijai Mohan , Srikanth Thirumalai
CPC classification number: G06K9/46 , G06K9/00201 , G06T1/0007 , G06T7/0081 , G06T2207/10004
Abstract: Recommendations can be generated even in situations where sufficient user information is unavailable for providing personalized recommendations. Instead of generating recommendations for an item based on item type or category, a relation graph can be consulted that enables other items to be recommended that are related to the item in some way, which may be independent of the type or category of item. For example, images of models, celebrities, or everyday people wearing items of clothing, jewelry, handbags, shoes, and other such items can be received and analyzed to recognize those items and cause them to be linked in the relation graph. When generating recommendations or selecting advertisements, the relation graph can be consulted to recommend products that other people have obtained with the item from any of a number of sources, such that the recommendations may be more valuable to the user.
-
公开(公告)号：US10824940B1
公开(公告)日：2020-11-03
申请号：US15365722
申请日：2016-11-30
Applicant: Amazon Technologies, Inc.
Inventor： Oleg Rybakov , Siddharth Singh
Abstract: The present disclosure is directed to training, and providing recommendations via, a temporal ensemble of neural networks. The neural networks in the temporal ensemble can be trained at different times. For example, a neural network can be periodically trained using current item interaction data, for example once per day using purchase histories of users of an electronic commerce system. The item interaction data can be split into a more recent group and a less recent group, for example the last two weeks of data and the remainder of the last two years of data. The periodic training of neural networks, using updated data and the sliding windows created by the date split, results in a number of different models for predicting item interaction events. Using a collection of these neural networks together as a temporal ensemble can increase recommendation accuracy without requiring additional hardware for training.
-
公开(公告)号：US09697608B1
公开(公告)日：2017-07-04
申请号：US14301599
申请日：2014-06-11
Applicant: Amazon Technologies, Inc.
Inventor： Oleg Rybakov , Avinash Aghoram Ravichandran , Daniel Bibireata , Ajay Kumar Mishra , Wei Zhang
CPC classification number: G06T7/254 , G06K9/00456 , G06K9/00671 , G06K9/46 , G06K9/6211 , H04N7/18 , H04N7/183
Abstract: A computing device can be configured to analyze information, such as frames captured in a video by a camera in the computing device, to determine locations of objects in captured frames using a scene-based tracking approach without individually having to track the identified objects across the captured frames. The computing device can track scenes, a global planar surface, across newly captured frames and the changes to (or transformation) the scene can be used to determine updated locations for objects that were identified in previously captured frames. Changes to the scene between frames can be measured using various techniques for estimating homographies. An updated location for the particular object in the currently captured frame can be determined by adjusting the location of the object, as determined in the previously captured frame, with respect to the transformation of the scene between the previously captured frame and the currently captured frame.
-
公开(公告)号：US09600741B1
公开(公告)日：2017-03-21
申请号：US14661438
申请日：2015-03-18
Applicant: Amazon Technologies, Inc.
Inventor： Che-Chun Su , Vivek Shah , Oleg Rybakov
CPC classification number: G06K9/52 , G06T5/007 , G06T5/50 , G06T7/00 , G06T2207/10152 , G06T2207/20208 , G06T2207/20221 , H04N5/2355 , H04N5/2356 , H04N5/367
Abstract: A plurality of instances of image data can be analyzed, and favored aspects of each instance identified and utilized in generating an enhanced output image. For example, a plurality of instances of image data can be analyzed to identify metric values associated with each pixel location, such as contrast, saturation, and exposedness. A weight map corresponding to each metric is generated for each instance of image data, each weight map indicating a value for the metric at each pixel location of the instance of image data. The weight maps associated with each instance of image data are merged, and a Gaussian pyramid of the merged weight map for each instance of image data is determined along with a Laplacian pyramid for each instance of image data. The Gaussian pyramids and Laplacian pyramids are merged into a Laplacian pyramid, which is then collapsed to form an enhanced output image.
-
-
-
-
-
-
-
-
-
Search Results
16
records
(took 0.053 seconds)
