52results about How to "Maximizing similarity" patented technology

A system for providing content recommendations, including a frontend manager for receiving explicit events from a client application of a user and generating implicit events based upon additional user actions within the client application; a backend storage of data on events and users and an Online Data Store for the explicit events and the implicit events; a Data Processor for creating an explicit user model from the explicit events and an implicit user model from the implicit events; a pool of recommendation engines with one or more recommendation algorithms for receiving the explicit user model and assigning a ranked recommendation list of content items to the user as a result, and further including an aggregator controlled by the Data Processor for aggregating the ranked recommendation lists based on a user-dependent strategy, in order to obtain multiple content recommendation lists of ranked items to be delivered by the frontend server to the client application in a final arrangement, pull from the Online Data Store along with data on the content sources.

Provided are an apparatus, method, and medium for generating a panoramic image. In the apparatus, a previously captured image and a currently captured image are projected onto a predetermined curved surface so as to form a first mapping image and a second mapping image, respectively. Movement values in predetermined directions are calculated for maximizing a color similarity between overlapped sections of the first and second mapping images. A panoramic image is generated using the first and second mapping images and the calculated movement values. Therefore, overlapped sections of neighboring images can be precisely superposed and aligned when a panoramic image is generated using a series of images.

A device and a method for measuring visual similarity between two images. One image (Q) being referred to as the model and one image (T) being referred to as the target, the method comprises a prior step (E2) of segmenting the images into regions (Q_i, T_i), with each region there being associated at least one attribute (F) representative of at least one characteristic of the region. It furthermore comprises the steps of calculating (E3) the visual similarity between the pairs (Q_i, T_i) of possible regions of the two images (Q, T), by taking into account the distance (D(Q_i, T_i)) between the said attributes (F) of the regions (Q_i, T_i) matched and the areas of the regions (Q_i, T_i) matched; selecting (E4) a certain number of pairs (Q_i, T_i) of regions whose similarity is greater than a first fixed threshold (ε), calculating (E9) the global similarity between the two images, based on the pairs (Q_i, T_i) of regions selected.

A trailer hitch facilitating proper alignment and connection to the hitch coupler of a trailer. The trailer hitch includes a tow bar assembly and a throat having spaced parallel top and floor panels generally equal in spacing to a vertical thickness of the tow bar and rearwardly diverging side panels. The front of the throat is connected or connectable to an elongated tubular extension or receiver and is sized to be slidably carried within the throat assembly. A slide bar is carried between the tow bar and the throat and includes a stop which limits the distance the tow bar may be extended rearwardly from the throat and, when in the fully or a partially extended position, is movable side-to-side to facilitate proper coupling alignment between the hitch ball and a hitch coupling of a trailer.

A situation recognition apparatus includes: an optical information acquisition unit configured to acquire optical information; a storage configured to store a plurality of pieces of optical information; a processing unit configured to match a plurality of the pieces of optical information stored in the storage and optical information newly acquired by the optical information acquisition unit; and an output unit configured to output a result of the matching. The storage further stores a probabilistic model that numerically represents transitions between the plurality of pieces of optical information.

The present invention relates to a method for the registration and superimposition of image data when taking serial radiographs in medical imaging, wherein a plurality of image data sets for a region of a patient (17) that is being investigated are constructed at time intervals using an imaging system (1) and are referenced with a first image data set for the region that is being investigated that was constructed previously using said imaging system (1). In the above method, a location system (2) is used during the production of serial radiographs constantly, or at least at a respective proximity in time to the construction of individual data sets, to determine a current spatial position of the region being investigated in a reference system that is firmly connected to the imaging system (1), whereby in the construction of the first image data set, a first spatial position of the region that is being investigated is recorded. In the construction of some or all further image data sets, the respective current spatial position of the region that is being investigated is determined and an image content of each first image data set is geometrically adapted on the basis of the difference between the first and the current spatial position, such that compensation is made for a different spatial position of the region that is being investigated. The geometrically adapted first image data set or an image data set derived therefrom, or an image data set that is positionally connected thereto by registration is then displayed superimposed with the respective further image data set.

The present invention relates to a method for encoding and decoding an image signal and to corresponding apparatuses therefor. In particular, during the encoding and/or decoding of an image signal filtering with at least two filters is performed. The sequence of the filter application and possibly the filters are selected and the filtering is applied in the selected filtering order and with the selected filters. The determination of the sequence of applying the filters may be performed either separately in the same way at the encoder and at the decoder, or, it may be determined at the encoder and signaled to the decoder.

The present disclosure provides a method and apparatus for generating training data for human face recognition, a device and a computer storage medium, wherein the method comprises: inputting accessory-not-worn face images into a generative network, to obtain accessory-worn face images; using the accessory-worn face images as second training data for building the human face recognition model; wherein the generative network is a generative network in a generative adversarial network obtained by pre-training with first training data, the first training data including the accessory-not-worn face images and accessory-worn face images corresponding to a same user identifier. In the present disclosure, the accessory-worn face images obtained in a data augmentation manner greatly expand the amount of training data for building the human face recognition model, and thereby improve the recognition accuracy of the accessory-worn face images.

A method and apparatus of training an acoustic feature extracting model, a device and a computer storage medium. The method comprises: considering a first acoustic feature extracted respectively from speech data corresponding to user identifiers as training data; training an initial model based on a deep neural network based on a criterion of a minimum classification error, until a preset first stop condition is reached; using a triplet loss layer to replace a Softmax layer in the initial model to constitute an acoustic feature extracting model, and continuing to train the acoustic feature extracting model until a preset second stop condition is reached, the acoustic feature extracting model being used to output a second acoustic feature of the speech data; wherein the triplet loss layer is used to maximize similarity between the second acoustic features of the same user, and minimize similarity between the second acoustic features of different users.

Featured are methods and systems for reslicing-based nonrigid registration of multiparametric and modality images or objects. In such registration methods, global motion is modeled by multiple affine transforms (rigid), while local motion and slices matching is estimated using non-uniform radial reslicing of RTV to search for best match between reference volume slices and resampled target volume (RTV), orthogonally re-sliced to higher spatial resolution compared to the spatial resolution of the reference volume. The reference and target volumes can be in any plane and the method presented in this invention transfers them to target plane which can be any of axial, coronal, or sagittal and there is no need for the original and target scanning planes to be the same.

A computer-implemented method for clustering a plurality of observations included in a dataset. The method includes selecting a subset of variables from a set of variables, where each variable in the subset is associated with each observation in the dataset; for each of a first number of times, randomly selecting an initial value for each variable in the subset to generate a seed value set that includes an initial value for each variable in the subset; generating a number of pre-clusters equal to the first number based on a similarity between each observation in the dataset and each of the first number of seed value sets; and applying a hierarchical clustering algorithm to the first number of pre-clusters to derive a second number of clusters.

A spatial motion calculation apparatus includes an image position relation calculation unit that calculates first similarities based on an image position relation between an inputted groups of feature points 1 and 2, a spatial position relation calculation unit that calculates second similarities based on an spatial position relation between said inputted groups, a feature descriptor relation calculation unit that calculates third similarities based on a feature descriptor relation between said inputted groups, and a spatial motion calculation unit that estimates the spatial motion based on the result that integrates the first to third similarities.

A method for maximizing a repeatability between a base seismic survey and a monitor seismic survey of a same surveyed subsurface during a 4-dimensional (4D) project. The method includes receiving first seismic data associated with the base seismic survey; receiving second seismic data associated with the monitor seismic survey, wherein the monitor seismic survey is performed later in time than the base seismic survey; estimating subsurface reflection-points and incidence angles; determining 4D-binning based on the estimated subsurface reflection-points and incidence angles; and maximizing the repeatability between the first seismic data and the second seismic data by using the 4D-binning.

A method and apparatus of building an acoustic feature extracting model, and an acoustic feature extracting method and apparatus. The method of building an acoustic feature extracting model comprises: considering first acoustic features extracted respectively from speech data corresponding to user identifiers as training data; using the training data to train a deep neural network to obtain an acoustic feature extracting model; wherein a target of training the deep neural network is to maximize similarity between the same user's second acoustic features and minimize similarity between different users' second acoustic features. The acoustic feature extracting model according to the present disclosure can self-learn optimal acoustic features that achieves a training target. As compared with a conventional acoustic feature extracting manner with a preset feature type and transformation manner, the acoustic feature extracting manner of the present disclosure achieves better flexibility and higher accuracy.

Specifically, a method decodes a picture. The picture is encoded and represented by blocks in a bitstream. For each block, a remap flag is obtained from the bit-stream. The block is either a remapped reconstructed block or a non-remapped reconstructed block. Either the non-mapped reconstructed block or an inverse remapped reconstructed block is output according to the remap flag. The remapped reconstructed block maximizes a similarity with the neighboring blocks, as compared to the similarity of the non-mapped reconstructed block and the neighboring blocks, by applying point operations to the remapped reconstructed block.

The disclosed embodiments provide a system that generates sound primitives to facilitate sound recognition. First, the system performs a feature-detection operation on sound samples to detect a set of sound features, wherein each sound feature comprises a measurable characteristic of a window of consecutive sound samples. Next, the system creates feature vectors from coefficients generated by the feature-detection operation, wherein each feature vector comprises a set of coefficients for sound features detected in a window. The system then performs a clustering operation on the feature vectors to produce feature-vector clusters, wherein each feature-vector cluster comprises a set of feature vectors that are proximate to each other in a feature-vector space that contains the feature vectors. After the clustering operation, the system defines a set of sound primitives, wherein each sound primitive is associated with a feature-vector cluster. Finally, the system associates semantic labels with the set of sound primitives.

A method for increasing similarity between a base seismic survey and a monitor seismic survey of a same surveyed subsurface during a 4-dimensional (4D) project. The method includes receiving first seismic data associated with the base seismic survey; receiving second seismic data associated with the monitor seismic survey, wherein the monitor seismic survey is performed later in time than the base seismic survey; migrating the first and second seismic data to an image-domain; and calculating, with a processor, a set of decimating weights based on the migrated first and second seismic data in the image-domain, to maximize a similarity between the first seismic data and the second seismic data.

Individual health vulnerability is assessed by obtaining health risk prevalence level data containing health risk prevalence levels for one or more health risks over a given geographical area. The health risk prevalence level data to generate for each health risk a prevalence level forecast as a function of time and location. A health risk prevalence level map is generated for the given geographical area illustrating current and future health risk prevalence levels for each health risk at a plurality of locations within the given geographical area. Personal health status data are obtained for a given individual along with a proposed travel itinerary f covering at least a portion of the geographical area over a given time duration. The personal health status data, travel itinerary and map generate a personal health risk vulnerability model containing a quantification of vulnerability to the one or more health risks resulting from the travel itinerary.

The present invention relates to a large-range-first cross-camera visual target re-identification method. The method comprises: step S1, obtaining initial single-camera tracks of targets; step S2, calculating a piecewise major color spectrum histogram feature of each track, and obtaining a track feature representation; step S3, obtaining a calculation formula of the similarity between any two tracks by using a minimum uncertainty method, so as to obtain the similarity between any two tracks; and step S4, performing global data association on all the tracks by using a maximum posterior probability method, so as to obtain a cross-camera tracking result. The target re-identification method of the present invention achieves high correct identification accuracy.

Individual health vulnerability is assessed by obtaining health risk prevalence level data containing health risk prevalence levels for one or more health risks over a given geographical area. The health risk prevalence level data to generate for each health risk a prevalence level forecast as a function of time and location. A health risk prevalence level map is generated for the given geographical area illustrating current and future health risk prevalence levels for each health risk at a plurality of locations within the given geographical area. Personal health status data are obtained for a given individual along with a proposed travel itinerary f covering at least a portion of the geographical area over a given time duration. The personal health status data, travel itinerary and map generate a personal health risk vulnerability model containing a quantification of vulnerability to the one or more health risks resulting from the travel itinerary.

Individual health vulnerability is assessed by obtaining health risk prevalence level data containing health risk prevalence levels for one or more health risks over a given geographical area. The health risk prevalence level data to generate for each health risk a prevalence level forecast as a function of time and location. A health risk prevalence level map is generated for the given geographical area illustrating current and future health risk prevalence levels for each health risk at a plurality of locations within the given geographical area. Personal health status data are obtained for a given individual along with a proposed travel itinerary f covering at least a portion of the geographical area over a given time duration. The personal health status data, travel itinerary and map generate a personal health risk vulnerability model containing a quantification of vulnerability to the one or more health risks resulting from the travel itinerary.

Individual health vulnerability is assessed by obtaining health risk prevalence level data containing health risk prevalence levels for one or more health risks over a given geographical area. The health risk prevalence level data to generate for each health risk a prevalence level forecast as a function of time and location. A health risk prevalence level map is generated for the given geographical area illustrating current and future health risk prevalence levels for each health risk at a plurality of locations within the given geographical area. Personal health status data are obtained for a given individual along with a proposed travel itinerary f covering at least a portion of the geographical area over a given time duration. The personal health status data, travel itinerary and map generate a personal health risk vulnerability model containing a quantification of vulnerability to the one or more health risks resulting from the travel itinerary.