Microsoft Technology Licensing, LLC

Beijing Nmc Co., Ltd.

Bo Geng

Xian-Sheng Hua

Linjun Yang

Arun Sacheti

Pak-Ming Cheung

Xin Yu

Xu Bin Liu

Dengyong Zhou

Zhong Wu

Xuewei Wu

Guoqing Qiu

Yang Cai

Yangchao Li

Architecture that addresses page information lost as part of a selection process in a search engine framework. An aggregation process collects all page or document information from the same image cluster and uses the aggregated page information to annotate one or more selected image-page pairs within the same image cluster. Once the entire set of descriptive terms is received, the entire set of descriptive terms or only an optimum set of top N descriptive terms of the entire set is for annotation of one or more of the representative images in the cluster.

Image annotation using aggregated page information from active and inactive indices

Provided is a magnetron source, which comprises a target material, a magnetron located thereabove and a scanning mechanism connected to the magnetron for controlling the movement of the magnetron above the target material. The scanning mechanism comprises a peach-shaped track, with the magnetron movably disposed thereon; a first driving shaft, with the bottom end thereof connected with the origin of the polar coordinates of the peach-shaped track, for driving the peach-shaped track to rotate about the axis of the first driving shaft; a first driver connected to the first driving shaft for driving the first driving shaft to rotate; and a second driver for driving the magnetron to move along the peach-shaped track via a transmission assembly. A magnetron sputtering device including the magnetron and a method for magnetron sputtering using the magnetron sputtering device are also provided.

Magnetron source, magnetron sputtering apparatus and magnetron sputtering method

Architecture that performs image page index selection. A learning-based framework learns a statistical model based on the hyperlink (URL-uniform resource locator) previous click information obtained from the image search users. The learned model can combine the features of a newly discovered URL to predict the possibility of the newly-discovered URL being clicked in the future image search. In addition to existing web index selection features, image clicks are added as features, and the image clicks are aggregated over different URL segments, as well as the site modeling pattern trees to reduce the sparse problem of the image click information.

Learning-based image webpage index selection

This document describes techniques that utilize a learning method to generate a ranking model for use in image search systems. The techniques leverage textual information and visual information simultaneously when generating the ranking model. The tools are further configured to apply the ranking model responsive to receiving an image search query.

Content-aware ranking for visual search

Some implementations provide techniques and arrangements to perform image retrieval. For example, some implementations identify an object of interest and a visual context in a first image. In some implementations, a second image that includes a second object of interest and a second visual context may be compared to the object of interest and the visual content, respectively, to determine whether the second image matches the first image.

Object retrieval using visual query context

Techniques described herein create an accurate active-learning model that takes into account a sample selection bias of elements, such as images, selected for labeling by a user. These techniques select a first set of elements for labeling. Once a user labels these elements, the techniques calculate a sample selection bias of the selected elements and train a model that takes into account the sample selection bias. The techniques then select a second set of elements based, in part, on a sample selection bias of the elements. Again, once a user labels the second set of elements the techniques train the model while taking into account the calculated sample selection bias. Once the trained model satisfies a predefined stop condition, the techniques use the trained model to predict labels for the remaining unlabeled elements.

Growing community of inventors

Beijing, China

Bo Geng