Abstract: Methods, systems, and computer-readable media for search query extension are disclosed. A particular method retrieves a set of queries related to a search query from a query repository. The search query includes one or more search words and at least one search word is included in each query in the set of queries. A set of context words based on one or more computer-readable files accessible to a computer is generated and an intersection is calculated. The intersection is based on the set of queries and the set of context words. An extended search query is generated based on the calculated intersection and the search query.

Abstract: A computer may identify an individual according to one or more biometrics based on various physiological aspects of the individual, such as metrics of various features of the face, gait, fingerprint, or voice of the individual. However, biometrics are often computationally intensive to compute, inaccurate, and unable to scale to identify an individual among a large set of known individuals. Therefore, the biometric identification of an individual may be supplemented by identifying one or more devices associated with the individual (e.g., a mobile phone, a vehicle driven by the individual, or an implanted medical device). When an individual is registered for identification, various device identifiers of devices associated with the individual may be stored along with the biometrics of the individual. Individuals may then be identified using both biometrics and detected device identifiers, thereby improving the efficiency, speed, accuracy, and scalability of the identification.

Abstract: A training set for a post-filter classifier is created from the output of a face detector. The face detector can be a Viola Jones face detector. Face detectors produce false positives and true positives. The regions in the training set are labeled so that false positives are labeled negative and true positives are labeled positive. The labeled training set is used to train a post-filter classifier. The post-filter classifier can be an SVM (Support Vector Machine). The trained face detection classifier is placed at the end of a face detection pipeline comprising a face detector, one or more feature extractors and the trained post-filter classifier. The post-filter reduces the number of false positives in the face detector output while keeping the number of true positives almost unchanged using features different from the Haar features used by the face detector.

Abstract: In one embodiment, the present invention includes a method for training a Support Vector Machine (SVM) on a subset of features (d?) of a feature set having (d) features of a plurality of training instances to obtain a weight per instance, approximating a quality for the d features of the feature set using the weight per instance, ranking the d features of the feature set based on the approximated quality, and selecting a subset (q) of the features of the feature set based on the ranked approximated quality. Other embodiments are described and claimed.

Abstract: Briefly, a method an apparatus and a wireless communication device are provided. The wireless communication device includes a receiver to receive complex sequences of symbols. The receiver includes an estimator to estimate one or more channel taps. The estimator includes a memory to store at least a portion of one or more calculated values of an estimation matrix and is capable to estimate the one or more channel taps based on a stored portion of calculated values of the estimation matrix.

Abstract: In one embodiment, the present invention includes a method for training a Support Vector Machine (SVM) on a subset of features (d?) of a feature set having (d) features of a plurality of training instances to obtain a weight per instance, approximating a quality for the d features of the feature set using the weight per instance, ranking the d features of the feature set based on the approximated quality, and selecting a subset (q) of the features of the feature set based on the ranked approximated quality. Other embodiments are described and claimed.

Abstract: A social network application may identify images having common links between a first user's image collection and a second user's image collection. The common links may be identified through metadata or similar portions of the images. Using the first user's image collection, elements of interest may be identified and compared to a second user's image collection to find matches. When matches are found, the results may be selected from groups of results to show a diverse set of matches. The user may be presented with options to select and add matched images to the user's collection, as well as to browse more images that match one or more of the groups.

Abstract: A user's collection of images may be analyzed to identify people's faces within the images, then create clusters of similar faces, where each of the clusters may represent a person. The clusters may be ranked in order of size to determine a relative importance of the associated person to the user. The ranking may be used in many social networking applications to filter and present content that may be of interest to the user. In one use scenario, the clusters may be used to identify images from a second user's image collection, where the identified images may be pertinent or interesting to the first user. The ranking may also be a function of user interactions with the images, as well as other input not related to the images. The ranking may be incrementally updated when new images are added to the user's collection.

Abstract: Images from two image databases may be correlated based on identifying a common event, which may be determined by image metadata as well as image content. The image metadata may include timestamps, geotagging metadata, or other tags, as well as input from a social network application in some embodiments. The image content may include analysis to find common persons based on facial recognition or color histograms, common background components, or other common features. The common event may be used to identify images that may be shared among the participants of the event by a social network application, as well as other purposes.

Abstract: Systems and methods are described herein for finding and sharing digital images of a user, such as digital photographs of the user, that are located in collections of digital images belonging to others. In accordance with at least one implementation, a face model of a first user is built using a first user computer, wherein the face model is built based on digital images of the first user stored on or accessible to the first user computer. The face model of the first user is then made accessible to a second user computer for use by the second user computer in finding digital images of the first user stored on or accessible to the second user computer. The digital images found by the second user computer are then made accessible to the first user computer.

Abstract: An application management system may have a user interface in which a user may input a text phrase that describes a desired action. The system may generate metadata relating to the text phrase and distribute the metadata and text phrase to many different registered applications, some of which may be web based applications. Each application may return one or more suggested actions, along with some optional information from the application. The suggested actions may be ranked and presented on the user interface, and a user may select an action to be performed. The system may launch the application and have the action performed.

Abstract: Briefly, some embodiments of the invention may provide devices, systems and methods to estimate, detect and handle cross-correlation between co-channel interference and a received signal in a wireless communication network. For example, a method in accordance with an embodiment of the invention may handle a data block of a received signal based on an estimated cross-correlation between co-channel interference and the received signal.

Abstract: Methods, systems, and computer-readable media for search query extension are disclosed. A particular method retrieves a set of queries related to a search query from a query repository. The search query includes one or more search words and at least one search word is included in each query in the set of queries. A set of context words based on one or more computer-readable files accessible to a computer is generated and an intersection is calculated. The intersection is based on the set of queries and the set of context words. An extended search query is generated based on the calculated intersection and the search query.

Abstract: In one embodiment, the present invention includes a method for training a Support Vector Machine (SVM) on a subset of features (d?) of a feature set having (d) features of a plurality of training instances to obtain a weight per instance, approximating a quality for the d features of the feature set using the weight per instance, ranking the d features of the feature set based on the approximated quality, and selecting a subset (q) of the features of the feature set based on the ranked approximated quality. Other embodiments are described and claimed.

Abstract: A device, system and method for applying a whitening transformation to a plurality of samples of a sampled symbol to reduce both temporal correlation between the sampled symbol and a pre-determined number of previous symbols, and a spatial correlation between a plurality of samples.

Abstract: A method comprises receiving a wireless signal at a channel estimator, and receiving the wireless signal at a channel length determination unit. Each of the channel estimator and the channel length determination unit independently receive the wireless signal.

Abstract: A receiver estimates a channel parameter from a received signal. The channel parameter may a DC signal level, a carrier to interference ratio, a noise spectrum, or the like. The receiver performs channel equalization and produces estimated symbols. The estimated symbols are remodulated and the channel parameter is re-estimated. An initial estimation error is calculated as the difference between the two estimates. The estimated symbols are weighted using the initial estimation error, and channel decoding is performed. Estimated symbol weights may be reduced when the initial estimation error is above a threshold, and may be increased when the initial estimation error is below a threshold.

Abstract: A receiver includes a channel length determination unit to determine a channel length for use by a channel estimator. In at least one embodiment, an iterative process is used to find a channel length (or a combination of channel length and one or more other channel related parameters) that optimizes a channel estimation quality criterion. A time tracking method is also provided that uses an early metric, a center metric, and a late metric to determine an appropriate timing.

Abstract: An iterative equalization process uses a reduced complexity equalizer as a first stage in the equalization process and a reduced alphabet MLSE equalizer as a later stage. The reduced complexity equalizer is used to identify a reduced alphabet including symbols having a higher probability of representing an actual transmitted symbol. The reduced alphabet is then used as the alphabet for the reduced alphabet MLSE equalizer.

Abstract: A device, system and method for applying a whitening transformation to a plurality of samples of a sampled symbol to reduce both temporal correlation between the sampled symbol and a pre-determined number of previous symbols, and a spatial correlation between a plurality of samples.