公开(公告)号：US09098888B1

公开(公告)日：2015-08-04

申请号：US14105028

申请日：2013-12-12

Applicant: A9.com, Inc.

Inventor： Xiaofan Lin ,  Adam Wiggen Kraft ,  Yu Lou ,  Douglas Ryan Gray ,  Colin Jon Taylor

IPC: G06T7/00

CPC classification number: G06K9/18 ,  G06K9/00456 ,  G06K9/00523 ,  G06K9/228 ,  G06K2209/01 ,  G06T7/11

Abstract: Various embodiments provide methods and systems for identifying text in an image by applying suitable text detection parameters in text detection. The suitable text detection parameters can be determined based on parameter metric feedback from one or more text identification subtasks, such as text detection, text recognition, preprocessing, character set mapping, pattern matching and validation. In some embodiments, the image can be defined into one or more image regions by performing glyph detection on the image. Text detection parameters applying to each of the one or more image regions can be adjusted based on measured one or more parameter metrics in the respective image region.

Abstract translation: 各种实施例提供了通过在文本检测中应用合适的文本检测参数来识别图像中的文本的方法和系统。 可以基于来自一个或多个文本识别子任务的参数度量反馈（例如文本检测，文本识别，预处理，字符集映射，模式匹配和验证）来确定合适的文本检测参数。 在一些实施例中，可以通过对图像执行字形检测来将图像定义为一个或多个图像区域。 可以基于相应图像区域中测量的一个或多个参数度量来调整应用于一个或多个图像区域中的每一个的文本检测参数。

公开(公告)号：US11093748B2

公开(公告)日：2021-08-17

申请号：US16773763

申请日：2020-01-27

Applicant: A9.com, Inc.

Inventor： Peiqi Tang ,  Andrea Zehr ,  Rupa Chaturvedi ,  Yu Lou ,  Colin Jon Taylor ,  Mark Scott Waldo ,  Shaun Michael Post

IPC: G06K9/00 ,  G06T19/00 ,  G06K9/62 ,  G06T13/80 ,  G06K9/46 ,  G06F16/532 ,  G06F16/583 ,  G06T7/246 ,  G06K9/22 ,  G06K9/32

Abstract: Various embodiments of the present disclosure provide systems and method for visual search and augmented reality, in which an onscreen body of visual markers overlayed on the interface signals the current state of an image recognition process. Specifically, the body of visual markers may take on a plurality of behaviors, in which a particular behavior is indicative of a particular state. Thus, the user can tell what the current state of the scanning process is by the behavior of the body of visual markers. The behavior of the body of visual markers may also indicate to the user recommended actions that can be taken to improve the scanning condition or otherwise facilitate the process. In various embodiments, as the scanning process goes from one state to another state, the onscreen body of visual markers may move or seamlessly transition from one behavior to another behavior, accordingly.

公开(公告)号：US10706098B1

公开(公告)日：2020-07-07

申请号：US15084256

申请日：2016-03-29

Applicant: A9.com, Inc.

Inventor： Sunil Ramesh ,  Shruti Sheorey ,  Colin Jon Taylor

IPC: G06F16/58 ,  G06Q30/06 ,  G06F16/583 ,  G06F16/2457 ,  G06F40/134

Abstract: Techniques for providing recommended keywords in response to an image-based query are disclosed herein. In particular, various embodiments utilize an image matching service to identify recommended search keywords associated with image data received from a user. The search keywords can be used to perform a keyword search to identify content associated with an image input that may be relevant. For example, an image search query can be received from a user. The image search query may result in multiple different types of content that are associated with the image. The system may present keywords associated with matching images to allow a user to further refine their search and/or find other related products that may not match with the particular image. This enables users to quickly refine a search using keywords that may be difficult to identify otherwise and to find the most relevant content for the user.

公开(公告)号：US20190272425A1

公开(公告)日：2019-09-05

申请号：US15911850

申请日：2018-03-05

Applicant: A9.com, Inc.

Inventor： Peiqi Tang ,  Andrea Zehr ,  Rupa Chaturvedi ,  Yu Lou ,  Colin Jon Taylor ,  Mark Scott Waldo ,  Shaun Michael Post

IPC: G06K9/00 ,  G06T19/00 ,  G06K9/62 ,  G06K9/46 ,  G06T13/80 ,  G06F17/30

Abstract: Various embodiments of the present disclosure provide systems and method for visual search and augmented reality, in which an onscreen body of visual markers overlayed on the interface signals the current state of an image recognition process. Specifically, the body of visual markers may take on a plurality of behaviors, in which a particular behavior is indicative of a particular state. Thus, the user can tell what the current state of the scanning process is by the behavior of the body of visual markers. The behavior of the body of visual markers may also indicate to the user recommended actions that can be taken to improve the scanning condition or otherwise facilitate the process. In various embodiments, as the scanning process goes from one state to another state, the onscreen body of visual markers may move or seamlessly transition from one behavior to another behavior, accordingly.

公开(公告)号：US10210423B2

公开(公告)日：2019-02-19

申请号：US15166973

申请日：2016-05-27

Applicant: A9.com, Inc.

Inventor： Ismet Zeki Yalniz ,  Colin Jon Taylor ,  Mehmet Nejat Tek ,  Shanghsuan Tsai

IPC: G06K9/46 ,  G06K9/00 ,  G06T7/60 ,  G06K9/62 ,  G06T7/11 ,  G06T7/90

Abstract: Object identification through image matching can utilize ratio and other data to accurately identify objects having relatively few feature points otherwise useful for identifying objects. An initial image analysis attempts to locate a “scalar” in the image, such as may include a label, text, icon, or other identifier that can help to narrow a classification of the search, as well as to provide a frame of reference for relative measurements obtained from the image. By comparing the ratios of dimensions of the scalar with other dimensions of the object, it is possible to discriminate between objects containing that scalar in a way that is relatively robust to changes in viewpoint. A ratio signature can be generated for an object for use in matching, while in other embodiments a classification can identify priority ratios that can be used to more accurately identify objects in that classification.

公开(公告)号：US10140549B2

公开(公告)日：2018-11-27

申请号：US15443730

申请日：2017-02-27

Applicant: A9.com, Inc.

Inventor： Simant Dube ,  Sunil Ramesh ,  Xiaofan Lin ,  Arnab Sanat Kumar Dhua ,  Colin Jon Taylor ,  Jaishanker K. Pillai

IPC: G06K9/62 ,  G06F17/30 ,  G06K9/00 ,  G06K9/46 ,  H04N19/90 ,  H04N19/426 ,  G06K9/52 ,  G06K9/66

Abstract: Various embodiments may increase scalability of image representations stored in a database for use in image matching and retrieval. For example, a system providing image matching can obtain images of a number of inventory items, extract features from each image using a feature extraction algorithm, and transform the same into their feature descriptor representations. These feature descriptor representations can be subsequently stored and used to compare against query images submitted by users. Though the size of each feature descriptor representation isn't particularly large, the total number of these descriptors requires a substantial amount of storage space. Accordingly, feature descriptor representations are compressed to minimize storage and, in one example, machine learning can be used to compensate for information lost as a result of the compression.

公开(公告)号：US20180330544A1

公开(公告)日：2018-11-15

申请号：US15590889

申请日：2017-05-09

Applicant: A9.com, Inc.

Inventor： Nicholas Corso ,  Michael Patrick Cutter ,  Yu Lou ,  Sean Niu ,  Shaun Michael Post ,  Colin Jon Taylor ,  Mark Scott Waldo

IPC: G06T19/00 ,  G06K9/00 ,  G06Q30/06

CPC classification number: G06T19/006 ,  G06K9/00671 ,  G06Q30/0623 ,  G06Q30/0643

Abstract: Systems and methods for a markerless approach to displaying an image of a virtual object in an environment are described. A computing device is used to capture an image of a real-world environment; for example including a feature-rich planar surface. One or more virtual objects which do not exist in the real-world environment are displayed in the image, such as by being positioned in a manner that they appear to be resting on the planar surface, based at least on a sensor bias value and scale information obtained by capturing multiple image views of the real-world environment.

公开(公告)号：US10109051B1

公开(公告)日：2018-10-23

申请号：US15196644

申请日：2016-06-29

Applicant: A9.com, Inc.

Inventor： Aishwarya Natesh ,  Arnab Sanat Kumar Dhua ,  Ming Du ,  R. Manmatha ,  Colin Jon Taylor ,  Mehmet Nejat Tek

IPC: G06T7/00 ,  G06T7/40 ,  G06K9/46 ,  G06T7/60 ,  G06K9/52 ,  G06K9/62 ,  G06Q30/06

Abstract: Images may be analyzed to determine a visually cohesive color palette, for example by comparing a subset of the colors most frequently appearing in the image to a plurality of color schemes (e.g., complementary, analogous, etc.), and potentially modifying one or more of the subset of colors to more accurately fit the selected color scheme. Various regions of the image are selected and portions of the regions having one or more colors of the color palette are extracted and classified to generate and compare feature vectors of the patches to previously-determined feature vectors of items to identify visually similar items. The visually similar items are selected for presentation in various ways, such as by choosing an outfit of visually-similar apparel items based on the locations of the corresponding colors in the image, etc.

公开(公告)号：US20170263011A1

公开(公告)日：2017-09-14

申请号：US15606864

申请日：2017-05-26

Applicant: A9.com, Inc.

Inventor： Ismet Zeki Yalniz ,  Leo Benedict Baldwin ,  Arnab Sanat Kumar Dhua ,  Isabella Talley Lewis ,  Colin Jon Taylor

IPC: G06T7/42 ,  G06K9/62 ,  G06T7/45

CPC classification number: G06T7/42 ,  G06K9/00664 ,  G06K9/2027 ,  G06K9/3233 ,  G06K9/342 ,  G06K9/4604 ,  G06K9/6201 ,  G06T7/11 ,  G06T7/194 ,  G06T7/45 ,  G06T7/90 ,  G06T2207/10024

Abstract: The recognition of objects such as clothing items can be improved by capturing image data that represents the material characteristics and true color of the object. A high resolution image can be captured that shows the individual threads and/or fibers, among other possible pattern or texture characteristics. Various approaches are discussed for determining the scale of these features to more accurately determine the way the material is made. Further, approaches are discussed that enable the true color of the material to be determined, through hardware and/or software, to further improve the accuracy of such a determination. The material and color properties can be combined with a conventional object recognition process to provide accurate object identification and/or authentication results.

公开(公告)号：US09275682B1

公开(公告)日：2016-03-01

申请号：US14498818

申请日：2014-09-26

Applicant: A9.com, Inc.

Inventor： Ismet Zeki Yalniz ,  Adam Carlson ,  Douglas Ryan Gray ,  Colin Jon Taylor

IPC: G06K9/34 ,  G11B27/30 ,  G11B27/036

CPC classification number: G11B27/3072 ,  G11B27/031 ,  G11B27/036 ,  G11B27/10 ,  G11B27/3081 ,  G11B27/34

Abstract: Various embodiments identify differences between frame sequences of a video. For example, to determine a difference between two versions of a video, a fingerprint of each frame of the two versions is generated. From the fingerprints, a run-length encoded representation of each version is generated. The fingerprints which appear only once (i.e., unique fingerprints) in the entire video are identified from each version and compared to identify matching unique fingerprints across versions. The matching unique fingerprints are sorted and filtered to determine split points, which are used to align the two versions of the video. Accordingly, each version is segmented into smaller frame sequences using the split points. Once segmented, the individual frames of each segment are aligned across versions using a dynamic programming algorithm. After aligning the segments at a frame level, the segments are reassembled to generate a global alignment output.

Abstract translation: 各种实施例识别视频的帧序列之间的差异。 例如，为了确定视频的两个版本之间的差异，生成两个版本的每个帧的指纹。 从指纹中，生成每个版本的游程长度编码表示。 从每个版本识别整个视频中仅出现一次的指纹（即，唯一指纹），并进行比较以识别跨越版本的匹配的唯一指纹。 匹配的唯一指纹被分类和过滤以确定分割点，其用于对准视频的两个版本。 因此，使用分割点将每个版本分割成较小的帧序列。 一旦分段，每个段的各个帧在版本之间使用动态规划算法对齐。 在帧级别对齐段之后，重新组合段以产生全局对准输出。