Abstract: Attractiveness of an image may be estimated by integrating extracted visual features with contextual cues pertaining to the image. Image attractiveness may be defined by the visual features (e.g., perceptual quality, aesthetic sensitivity, and/or affective tone) of elements contained within the image. Images may be indexed based on the estimated attractiveness, search results may be presented based on image attractiveness, and/or a user may elect, after receiving image search results, to re-rank the image search results by attractiveness.

Abstract: A facility for visual search on a mobile device takes advantage of multi-modal and multi-touch input on the mobile device. By extracting lexical entities from a spoken search query and matching the lexical entities to image tags, the facility provides candidate images for each entity. Selected ones of the candidate images are used to construct a composite visual query image on a query canvas. The relative size and position of the selected candidate images in the composite visual query image, which need not be an existing image, contribute to a definition of a context of the composite visual query image being submitted for context-aware visual search.

Abstract: The techniques described herein determine an initial set of ranked images associated with an image-based search query. Based on visual content similarities between images in the initial set of ranked images, the techniques select confident image samples from the initial set of ranked images. The techniques then use the confident image samples to rerank the initial set of ranked images. Accordingly, a search engine uses the confident image samples to promote images that are likely to be relevant to the search query, while demoting images that are not likely to be relevant to the search query. Therefore, the search engine can provide improved relevance-based search results to an image-based search query.

Abstract: Some examples include receiving a microblog entry from a social stream domain. Further, some implementations include determining, based on a topic space associated with the social stream domain and a media domain, a topic that is associated with the microblog entry. Some implementations include determining, based on the topic space, one or more media items that are associated with the topic.

Abstract: The present motion description technique provides a technique for defining a motion description offline. The motion description can then later be extracted from a multimedia representation and adapted to various multimedia-related applications in a manner that not only reduces the processing for motion estimation but also provides high compression performance during an encoding/transcoding process. The motion description technique employs a motion alignment scheme utilizing a hierarchical model to describe motion data of each macroblock in a coarse-to-fine manner. Motion information is obtained for motion vectors of macroblocks for different partition modes. The resulting motion information is compressed based on correlations among spatially neighboring macroblocks and among partition modes to form the offline motion description.

Abstract: Visual concepts contained within a video clip are classified based upon a set of target concepts. The clip is segmented into shots and a multi-layer multi-instance (MLMI) structured metadata representation of each shot is constructed. A set of pre-generated trained models of the target concepts is validated using a set of training shots. An MLMI kernel is recursively generated which models the MLMI structured metadata representation of each shot by comparing prescribed pairs of shots. The MLMI kernel is subsequently utilized to generate a learned objective decision function which learns a classifier for determining if a particular shot (that is not in the set of training shots) contains instances of the target concepts. A regularization framework can also be utilized in conjunction with the MLMI kernel to generate modified learned objective decision functions. The regularization framework introduces explicit constraints which serve to maximize the precision of the classifier.

Abstract: Systems and methods provide image compression based on parameter-assisted inpainting. In one implementation of an encoder, an image is partitioned into blocks and the blocks classified as smooth or unsmooth, based on the degree of visual edge content and chromatic variation in each block. Image content of the unsmooth blocks is compressed, while image content of the smooth blocks is summarized by parameters, but not compressed. The parameters, once obtained, may also be compressed. At a decoder, the compressed image content of the unsmooth blocks and the compressed parameters of the smooth blocks are each decompressed. Each smooth block is then reconstructed by inpainting, guided by the parameters in order to impart visual detail from the original image that cannot be implied from the image content of neighboring blocks that have been decoded.

Abstract: Methods and apparatuses are provided for achieving improved video coding efficiency through the use of Motion Vector Predictors (MVPs) for the encoding or decoding of motion parameters within the calculation of the motion information in B pictures and/or P pictures. Certain exemplary methods and apparatuses selectively apply temporal and/or spatial prediction. Rate Distortion Optimization (RDO) techniques are also applied in certain methods and apparatuses to further help improve coding efficiency.

Abstract: A user behavior model provides personalized recommendations based in part on time and location, particularly to users of mobile devices. Entity types are ranked according to relevance to the user. Example entity types are restaurant, hotel, etc. The relevance may be based on reference to a large-scale database containing queries from other users. Additionally, entities within each entity type may be ranked based on relevance to the user and the time and location context. A user interface may display a ranked list of entity types, such as restaurant, hotel, etc., wherein each entity type is represented by a highest-ranked entity with the entity type. Thus, the user interface may display a highest-ranked restaurant, a highest-ranked hotel, etc. Upon user selection of one such entity type the user interface is replaced with a second user interface, for example showing a ranked hierarchy of restaurants, headed by the highest-ranked restaurant.

Abstract: A home cloud computing system employs a virtualization system to virtualize data of a device and adaptively transform type or format of the virtualized data for one or more other devices, thus leveraging resources of the device for the one or more other devices. Through data virtualization and adaptive transformation, devices of heterogeneous types are seamlessly connected to one another and can act as input or output devices for each other to create a home cloud network of devices.

Abstract: Some implementations include techniques and arrangements to provide document-related representative information with search results. For example, a representative query and/or representative results may be provided for one or more individual documents identified in a set of search results to supplement the search results returned in response to a received search query. The representative queries may be determined by correlating a plurality of previously submitted queries in search log data with a plurality of documents returned in response to the queries. In some implementations, click-through frequency for a particular document with respect to the plurality of queries may be taken into consideration when determining the representative queries for the particular document.

Abstract: This disclosure describes techniques for searching for similar images to an image query by using an approximate k-Nearest Neighbor (k-NN) graph. The approximate k-NN graph is constructed from data points partitioned into subsets to further identify nearest-neighboring data points for each data point. The data points may connect with the nearest-neighboring data points in a subset to form an approximate neighborhood subgraph. These subgraphs from all the subsets are combined together to form a base approximate k-NN graph. Then by performing more random hierarchical partition, more base approximate k-NN graphs are formed, and further combined together to create an approximate k-NN graph. The approximate k-NN graph expands into other neighborhoods and identifies the best k-NN data points. The approximate k-NN graph retrieves the best NN data points, based at least in part on the retrieved best k-NN data points representing images being similar in appearance to the image query.

Abstract: Techniques describe providing a web page for a proxy-based browser on a mobile device to enhance user experience. A proxy server receives a layout of the web page, extracts web elements from the web page, and captures images of the web elements of the web page. The web elements are incorporated with a background screen image to form a composite screen format to represent a display of the web page. The background screen image is compressed by splitting an encoded frame into fixed-size slices and splitting a previous screen frame into fixed-size slices. The proxy server provides the web page synchronized with the mobile device based on the composite screen format and the compressed background screen image. Furthermore, the proxy server receives input from a user to provide updates to web elements that are dynamic on the web page to be displayed on the screen of the mobile device.

Abstract: Improved methods and apparatuses are provided for switching of streaming data bitstreams, such as, for example, used in video streaming and other related applications. Some desired functionalities provided herein include random access, fast forward and fast backward, error-resilience and bandwidth adaptation. The improved methods and apparatuses can be configured to increase coding efficiency of and/or reduce the amount of data needed to encode a switching bitstream.

Abstract: An image-based text extraction and searching system extracts an image be selected by gesture input by a user and the associated image data and proximate textual data in response to the image selection. Extracted image data and textual data can be utilized to perform or enhance a computerized search. The system can determine one or more database search terms based on the textual data and generate at least a first search query proposal related to the image data and the textual data.

Abstract: Described herein is technology for, among other things, selecting a representative thumbnail from a video clip. The technology involves analyzing frames of the video clip to determine which frames are stable, the result of the analysis being a number of segments of stable frames. From the stable segments, a number of candidate segments are selected, where candidate segments are those segments determined to a degree of certainty to be program content. The representative thumbnail is then selected from among the frames of the candidate segments.

Abstract: Techniques for setting depth values for invalid measurement regions of depth images are described herein. A computing device may set the depth values based on evaluations of depth values of neighboring pixels and of corresponding pixels from time-adjacent depth images. Alternately or additionally, the computing device may utilize a texture image corresponding to the depth image to identify objects and may set depth values for pixels based on depth values of other pixels belonging to the same object. After setting the depth values, the computing device may normalize the depth values of the pixels. Further, the computing device may generate reduced representations of the depth images based on a depth reference model or a depth error model and may provide the reduced representations to an encoder.

Abstract: A proxy-based thin-client web browsing framework enables cooperative web browsing of multiple devices. The multiple devices may include devices that are not intended for web browsing and have limited or no web browsers and/or user input capabilities. The proxy-based thin client web browsing framework employs a virtual browser at a proxy server to perform all browser-engine logics, and retrieve, render and encode web pages on behalf of the multiple devices. The multiple devices therefore only need to have limited decoding and display capabilities to perform web browsing. The proxy-based thin client web browsing framework further includes a touch controller as a remote controller for a device that has no or limited user texting or manipulating capabilities.

Abstract: An “Interactive Virtual Display,” as described herein, provides various systems and techniques that facilitate ubiquitous user interaction with both local and remote heterogeneous computing devices. More specifically, the Interactive Virtual Display uses various combinations of small-size programmable hardware and portable or wearable sensors to enable any display surface (e.g., computer display devices, televisions, projected images/video from projection devices, etc.) to act as a thin client for users to interact with a plurality heterogeneous computing devices regardless of where those devices are located relative to the user. The Interactive Virtual Display provides a flexible system architecture that enables communication and collaboration between a plurality of both local and remote heterogeneous computing devices.

Abstract: Systems and methods for determining insertion points in a first video stream are described. The insertions points being configured for inserting at least one second video into the first video. In accordance with one embodiment, a method for determining the insertion points includes parsing the first video into a plurality of shots. The plurality of shots includes one or more shot boundaries. The method then determines one or more insertion points by balancing a discontinuity metric and an attractiveness metric of each shot boundary.