公开(公告)号：US09940575B2

公开(公告)日：2018-04-10

申请号：US14730476

申请日：2015-06-04

Applicant: Yahoo!, Inc.

Inventor： JenHao Hsiao

IPC: G06K9/66 ,  G06N3/08 ,  G06F17/30

CPC classification number: G06N3/08 ,  G06F17/30256 ,  G06N3/0427 ,  G06N3/0454 ,  G06N3/084

Abstract: As provided herein, a domain model, corresponding to a domain of an image, may be merged with a pre-trained fundamental model to generate a trained fundamental model. The trained fundamental model may comprise a feature description of the image converted into a binary code. Responsive to a user submitting a search query, a coarse image search may be performed, using a search query binary code derived from the search query, to identify a candidate group, comprising one or more images, having binary codes corresponding to the search query binary code. A fine image search may be performed on the candidate group utilizing a search query feature description derived from the search query. The fine image search may be used to rank images within the candidate group based upon a similarity between the search query feature description and feature descriptions of the one or more images within the candidate group.

公开(公告)号：US20160342623A1

公开(公告)日：2016-11-24

申请号：US14715246

申请日：2015-05-18

Applicant: Yahoo! Inc.

Inventor： JenHao Hsiao ,  Jia Li

IPC: G06F17/30

CPC classification number: G06F16/583

Abstract: An approach for performing mobile visual search uses deep variant coding of images to reduce the amount of data transmitted from mobile devices to a search server and to provide more efficient indexing and searching on the search server. The amount of data used to represent an image varies depending upon the content of the image and is less than conventional fixed bit length hashing approaches. Denser regions of a feature space are represented by more encoding bits and sparser regions of the feature space are represented by fewer encoding bits, so that the overall number of encoding bits for an image feature is reduced. The approach generally involves determining a set of hash functions that provide deep hashing with more evenly-distributed hash buckets. One or more additional hash functions may be selectively generated for particular hash buckets that contain more than a specified number of images.

Abstract translation: 用于执行移动视觉搜索的方法使用图像的深度变体编码来减少从移动设备传输到搜索服务器的数据量，并在搜索服务器上提供更有效的索引和搜索。 用于表示图像的数据量根据图像的内容而变化，并且小于常规固定位长度哈希方法。 特征空间的密集区域由更多的编码比特表示，并且特征空间的稀疏区域由较少的编码比特表示，使得图像特征的编码比特的总数减少。 该方法通常涉及确定一组哈希函数，它们使用更均匀分布的哈希桶来提供深度哈希。 可以针对包含多于指定数目的图像的特定哈希桶选择性地生成一个或多个附加散列函数。

公开(公告)号：US20160357748A1

公开(公告)日：2016-12-08

申请号：US14730476

申请日：2015-06-04

Applicant: Yahoo!, Inc.

Inventor： JenHao Hsiao

IPC: G06F17/30 ,  G06N3/08

CPC classification number: G06N3/08 ,  G06F17/30256 ,  G06N3/0427 ,  G06N3/0454 ,  G06N3/084

Abstract: As provided herein, a domain model, corresponding to a domain of an image, may be merged with a pre-trained fundamental model to generate a trained fundamental model. The trained fundamental model may comprise a feature description of the image converted into a binary code. Responsive to a user submitting a search query, a coarse image search may be performed, using a search query binary code derived from the search query, to identify a candidate group, comprising one or more images, having binary codes corresponding to the search query binary code. A fine image search may be performed on the candidate group utilizing a search query feature description derived from the search query. The fine image search may be used to rank images within the candidate group based upon a similarity between the search query feature description and feature descriptions of the one or more images within the candidate group.

Abstract translation: 如本文所提供的，对应于图像域的域模型可以与预训练的基本模型合并以产生经过训练的基本模型。 经过训练的基本模型可以包括转换成二进制码的图像的特征描述。 响应于提交搜索查询的用户，可以使用从搜索查询导出的搜索查询二进制代码来执行粗略图像搜索，以识别具有与搜索查询二进制对应的二进制代码的一个或多个图像的候选组 码。 可以利用从搜索查询导出的搜索查询特征描述对候选组执行精细图像搜索。 精细图像搜索可以用于基于候选组中的一个或多个图像的搜索查询特征描述和特征描述之间的相似度来对候选组内的图像进行排序。

公开(公告)号：US20160275111A1

公开(公告)日：2016-09-22

申请号：US14664011

申请日：2015-03-20

Applicant: Yahoo!, Inc.

Inventor： JenHao Hsiao

IPC: G06F17/30

CPC classification number: G06F17/30277 ,  G06F17/30256

Abstract: Users may have a variety of photos, but lack a mechanism to organize the photos. For example, a user may desire to access a photo of a child in front of a national monument, but may be unable to locate the photo amongst the photos. Accordingly, a photo query may be generated utilizing a photo query interface populated with a face query attribute (e.g., a proxy icon, a photo of a face, etc.), a face position attribute (e.g., a position of the face query attribute), and/or a location query attribute (e.g., a background of a photo, such as a forest, a monument, etc.) using drag and drop functionality and/or any other query construction functionality. One or more photos having attributes corresponding to the photo query may be identified and provided to the user.

Abstract translation: 用户可能会有各种各样的照片，但缺乏组织照片的机制。 例如，用户可能希望访问国家纪念碑前面的孩子的照片，但是可能无法在照片中定位照片。 因此，可以利用填充有面部查询属性（例如，代理图标，面部照片等）的照片查询界面，面部位置属性（例如，面部查询属性的位置）生成照片查询 ）和/或位置查询属性（例如，照片的背景，例如森林，纪念碑等）使用拖放功能和/或任何其他查询构造功能。 可以识别具有与照片查询相对应的属性的一个或多个照片，并将其提供给用户。

公开(公告)号：US20160005097A1

公开(公告)日：2016-01-07

申请号：US14325192

申请日：2014-07-07

Applicant: Yahoo! Inc.

Inventor： JenHao Hsiao ,  Jia Li

IPC: G06Q30/06

CPC classification number: G06Q30/0631

Abstract: Example methods, apparatuses, and/or articles of manufacture are disclosed that may be implemented, in whole or in part, using one or more computing devices to facilitate or support one or more processes and/or operations for one or more on-line recommendations, such as product-related recommendations, for example.

Abstract translation: 公开了可以全部或部分地使用一个或多个计算设备实现一个或多个在线建议的一个或多个过程和/或操作的一个或多个处理和/或操作的示例性方法，设备和/或制品 ，例如与产品相关的建议。

公开(公告)号：US20150220950A1

公开(公告)日：2015-08-06

申请号：US14174399

申请日：2014-02-06

Applicant: YAHOO! INC.

Inventor： JenHao Hsiao

IPC: G06Q30/02 ,  G06N99/00

CPC classification number: G06Q30/0203 ,  G06N20/00

Abstract: A relative labeling approach is disclosed to learn an item preference scoring function to rank items for a user. An iterative process may be used to present a set of items to a user in an interactive user interface, using which the user is asked to identify one of the items in the set that the user prefers over the other items in the set. Input received from the user may be considered to be a “labeling” of the items in the set relative to each other. Subsequent labeling input may be added to previous labeling input to generate an updated preference scoring function for the user. Selection of each item for inclusion in the set of items presented to the user may be based on a measure of the knowledge that may be gained by including the item in the set.

Abstract translation: 公开了一种相对标签方法来学习项目偏好评分功能以对用户排列项目。 迭代过程可以用于在交互式用户界面中向用户呈现一组项目，使用该组项目向用户标识用户偏爱集合中的其他项目的组中的一个项目。 从用户接收的输入可以被认为是相对于彼此的集合中的项目的“标注”。 随后的标签输入可以被添加到先前的标签输入以为用户生成更新的偏好评分功能。 每个项目的选择被包括在呈现给用户的一组项目中可以基于通过将该项目包括在该集合中可获得的知识的度量。