31 results about "Eigenvalue assignment" patented technology

Techniques are described herein for automatically evaluating the quality of digital images based on one or more color characteristics of the images. In some embodiments, a quality metric that indicates the likelihood that the digital images convey semantics is generated based on color characteristics of the digital images. The quality metric may be used, for example, to determine which keyframe to use to make a thumbnail to represent video data. In some embodiments, feature values are generated for an image based on color characteristics of the image, and the feature values are assigned to bins. In such embodiments, the quality metric may be generated to indicate how uniform the distribution of feature values is among the bins.

A method and apparatus for generating soft decision error correction code information. The method includes generating or creating a lookup table (LUT), such as a log likelihood ratio (LLR) lookup table, characterizing a flash memory device. The method also includes loading the lookup table into the SSD controller. The method also includes accessing the lookup table to assign LLR or other characteristic values to the cells of a flash memory device. The method also includes decoding the data in a flash memory device using the soft decision information provided by the lookup table and assigned to the appropriate cells of the flash memory device.

A method and apparatus for situation-based clustering digital photos, and a digital photo albuming system and method using the same. A situation-based digital photo clustering method of clustering digital photos based on a situation when a photo is taken includes: extracting photographing data information including at least a photographing time feature value from a digital photo file and extracting a content-based feature value from contents of a digital photo of the digital photo file; assigning an importance degree to each extracted photographing time feature value and content-based feature value and combining the values; and hierarchically clustering photographing situations using feature value information, the feature value information being the extracted photographing time feature value and content-based feature value combined with respect to the assigned degrees of importance.

A multimedia information arranging apparatus that utilizes various feature values contained in multimedia information, carries out a flexible retrieval and classification and displays the information in an information set arrangement space is provided.An information set obtaining portion 10 obtains information sets. An axis setting portion 20 assigns the feature value of the media information to an information set arrangement space axis so as to set the information set arrangement space. A feature value extracting portion 30 extracts the feature value from each of the media information of the obtained information set, and an information set arranging portion 40 arranges the information sets in the information set arrangement space according to their feature values. An information displaying portion 50 displays the information set arrangement space in such a manner as to be seen from a predetermined display viewpoint. By seeing the display result, as necessary, the axis setting portion 20 resets the information set arrangement space axis using a different feature value, so that the information sets are rearranged and redisplayed in the information set arrangement space, and retrievals of different aspects are repeated.

The invention discloses a turbo generator rotor interturn short circuit fault diagnosis method based on multi sensor joint. According to the method, a variety of fault characteristic parameters acquired by different types of sensors arranged at a number of parts of a generator and respective thresholds are used to carry out calculation, so as to acquire fault characteristic values when diagnosis is carried out according to each fault characteristic parameter; according to the reliability of different diagnostic methods, weight assigning is carried out on each fault characteristic value; the weighted sum of a variety of fault characteristic values is acquired; and finally by comparing the weighted sum with a comprehensive threshold, whether the turbo generator has a rotor winding interturn short circuit fault is identified. According to the invention, comprehensive analysis is carried out on the diagnosis information acquired by different types of sensors to judge whether the turbo generator has the rotor winding interturn short circuit fault; the method overcomes the shortcomings of the traditional single diagnosis method; and the sensitivity and the reliability of the rotor winding interturn short circuit fault detection of the turbo generator are improved.

A video action detection system uses feature a data extractor to extract feature data from video data at detected spatiotemporal interest points. A feature data quantizer assigns the extracted feature values to bins of a feature vector. Bin values are computed from a sum of contributions of spatiotemporal points of interest that have been assigned to the bin, with a bin dependent adjustment of a size of the sum and / or the contributions. The video action detection system computes a sum of match scores between the feature vector and reference vectors for the predetermined type of action. The bin dependent adjustment of the size of the sum and / or the contributions is adapted in a training step, based on partial match scores for individual bins obtained using an initial action detector.

A massive image library retrieving method based on an optimal K mean value Hash algorithm comprises the steps that part of images are selected from an image library to be retrieved to serve as a training image set, and firstly GIST characteristics of images of the training image set are extracted; characteristic value allocation preprocessing is conducted on characteristic data of the training image set; the preprocessed characteristic data are divided into a plurality of sub-spaces; a codebook and codes of the codebook of the corresponding sub-space are trained out for each sub-space; the processing and training process of characteristic data in the image library to be retrieved corresponds to the processing and training process of characteristic data in inquiring images, the GIST characteristics of images to be retrieved and the GIST characteristics of the inquiring images are extracted respectively, then Hash codes of the characteristics of the images to be retrieved and Hash codes of the characteristics of the inquiring images are calculated, the Hamming distance between the codes of the characteristics of the images to be retrieved and the codes of the characteristics of the inquiring images is calculated, and thus similar images are fast retrieved. The massive image library retrieving method based on the optimal K mean value Hash algorithm has good universality, the storage space for data is reduced, and the inquiring retrieving efficiency is improved.

There is disclosed an echo cancelling apparatus in a digital mobile communication system and method thereof, which is incorporated in a single chip together with a speech coder, and can improve convergence rate by using an excited signal as an input to n adaptive filter and also obtain a stabilized convergence characteristic while using a NLMS algorithm. The echo canceller according to the present invention uses the excited signal obtained at the speech coder as an input signal to an adaptive filter thereby to lower the Eigenvalues distribution rate of the speech signal, thus providing an improved performance compared to the conventional echo canceller.

The invention provides a method for user identification. The method comprises the following steps of receiving a user login request, wherein the user login request comprises more than one item of characteristic data of a user; computing a characteristic value corresponding to each item of characteristic data according to a preset function corresponding to each item of characteristic data in more than one item of characteristic data; computing a threshold corresponding to each item of characteristic data according to pre-stored sample data, wherein the threshold is used for distinguishing a characteristic value of a normal account and a characteristic value of an abnormal account; distributing a coefficient to the characteristic value corresponding to each item of characteristic data according to the threshold and the characteristic value of each item of characteristic data; and computing a first risk value corresponding to the user according to the characteristic value corresponding to each item of characteristic data, the coefficient distributed by the characteristic value corresponding to each item of characteristic data and a preset weight value corresponding to each item of characteristic data. The invention also provides a device for user identification. According to the method and the device, the risk of the user can be effectively identified, and the accuracy rate is high.

The invention discloses a method and a device for resource optimization of a broadband wireless access system. The method comprises the following steps: a base station receives data transmitted by a terminal and calculates the characteristic values of a channel between the base station and the terminal; and the base station transmits the power of each data flow to the terminal according to the limitation of system power and the allocation of the characteristic value of the channel. The method and the device can reduce the complexity of the algorithm, optimize the resource allocation, improve the system throughput and enhance the reliability of the transmission data link.

Systems, methods, and computer program products for autonomous car-like ground vehicle guidance and trajectory tracking control. A multi-loop 3DOF trajectory linearization controller (300) provides guidance to a vehicle having nonlinear rigid-body dynamics (332) with nonlinear tire traction force, nonlinear drag forces and actuator dynamics. The controller (300) may be based on a closed-loop PD-eigenvalue assignment and a singular perturbation (time-scale separation) theory for exponential stability, and controls the longitudinal velocity (Vx,sen) and steering angle (delta) simultaneously to follow a feasible guidance trajectory. A line-of-sight based pure-pursuit guidance controller (502) may generate a 3DOF spatial trajectory that is provided to the 3DOF controller (300) to enable targetpursuit and path-following / trajectory-tracking. The resulting combination may provide a 3DOF motion control system (500) with integrated simultaneous steering and speed control for automobile and car-like mobile robot target pursuit and trajectory-tracking.

The invention provides a method and a device for sample distributed clustering calculation. The method comprises that characteristic values of all the samples to be clustered are acquired to constitute a characteristic value set; a calculation speed of each kind of usable calculation equipment is estimated; and before a similarity degree between any two characteristic values in the characteristic value set is less than a preset threshold value, the following steps are repeated, wherein all the characteristic values in the characteristic value set are distributed to at least one kind of calculation equipment according to the calculation speed of each usable calculation equipment, so that the at least one kind of calculation equipment can screen the distributed characteristic values on the premise that processing time satisfies a preset condition, and the similarity degree between any two characteristic values is less than the preset threshold value. The method and the device for the sample distributed clustering calculation provided by the invention can solve the problem that existing clustering algorithms can hardly deal with a large scale of data, can realize the distributed clustering calculation of the large scale of data which can hardly be handled by the prior art, and can effectively increase efficiency of the clustering calculation.

Embodiments of the disclosure relate to decision tree generation based on pre-statistics of data for nodes. A plurality of data samples for nodes of a decision tree are obtained, the data samples having respective feature values with respect to a first feature. A target range is determined from a plurality of predefined ranges of values such that the number of feature values falling within the target range exceeds a predetermined threshold number. Then, the remaining feature values other than the feature values falling within the target range are assigned to the corresponding range of values,and the feature values falling into all the ranges of values are counted based on the allocation of the remaining feature values so as to be used for the allocation of the plurality of data samples tochild nodes of the nodes. In this way, the data processing speed and efficiency are significantly improved, and therefore the generation speed and efficiency of the decision tree are increased.

The invention provides an image processing method and a device, wherein, the method comprises the following steps: extracting X amplitude images of an image to be processed, fusing a plurality of amplitude images in different directions on the same scale to obtain Y fused amplitude images; Encoding the Y fused amplitude images, According to the local characteristics of each amplitude image, assigning each eigenvalue generated after encoding with weights, obtaining Y statistical histograms with weight information corresponding to the Y fused amplitude images, and cascading the Y statistical histograms to form feature vectors of the image to be processed. The technical scheme provided by the application is applicable to image recognition in a complex environment, and a more universal image recognition mechanism is realized.

The invention relates to a high-resolution remote sensing image city function partitioning method based on multi-feature fusion. The method comprises the steps of 1, image preprocessing; step 2, distributing the feature values in each image to the visual word most similar to the feature value, and counting the corresponding word frequency of each visual word to form a visual word feature; constructing a multi-feature BoW visual dictionary; 3, constructing an LDA probability topic model, and mining a high-dimensional semantic vector of the image by utilizing the LDA probability topic model; 4,training an SVM classifier according to the high-dimensional semantic vector obtained in the step 3; and step 5, carrying out urban function partitioning on the test set by using an SVM classifier. The method has the beneficial effects that the POI data is introduced, so that the wrong score of the remote sensing data caused by metamerism and foreign matters in the same spectrum is reduced; multiple features of the image, which comprises local features, spectral features, texture features, surface temperature features, spatial three-dimensional features and POI features, are comprehensively utilized, and higher classification precision can be obtained under the condition that the single feature of the image is not obvious.

The invention discloses an eigenvalue storage and retrieval method and device. The method comprises: extracting eigenvalues of images to obtain eigenvalues; judging whether the file to be written withthe eigenvalue is full or not, if not, writing the eigenvalue into the file; if yes, creating a new file, and writing the eigenvalue into the new file; assigning said eigenvalues search identifiers in the order in which they are written. In the embodiment of that present invention, after the corresponding eigenvalue of the image is determined, the eigenvalue is written into a file, and a search mark is assigned to the eigenvalue according to the order in which the eigenvalue is written. When the eigenvalue is retrieved, the search mark range information can be input according to the user's need, so that the corresponding file can be retrieved. The eigenvalue in the file is the retrieved eigenvalue, and the batch retrieval of the eigenvalue can be realized. Therefore, it is suitable for batch storage and retrieval of dynamic and static library scenarios.

The invention discloses a real-time super-resolution method and system based on an FPGA, and relates to the technical field of image processing. The method comprises the steps of splitting each frameof image in a medium with a relatively low resolution into sub-images for performing super-resolution processing; allocating a processing module according to an eigenvalue of the current frame of thesub-image; if the eigenvalue is relatively high, performing calculation by a neural network module; and if the eigenvalue is relatively low, performing calculation by an interpolation module. An output buffer outputs the high-resolution sub-images obtained by calculation and recombines the sub-images into a high-resolution image, so that a user can see the real-time ultra-high-definition image.

The invention discloses a turbo generator rotor interturn short circuit fault diagnosis method based on multi sensor joint. According to the method, a variety of fault characteristic parameters acquired by different types of sensors arranged at a number of parts of a generator and respective thresholds are used to carry out calculation, so as to acquire fault characteristic values when diagnosis is carried out according to each fault characteristic parameter; according to the reliability of different diagnostic methods, weight assigning is carried out on each fault characteristic value; the weighted sum of a variety of fault characteristic values is acquired; and finally by comparing the weighted sum with a comprehensive threshold, whether the turbo generator has a rotor winding interturn short circuit fault is identified. According to the invention, comprehensive analysis is carried out on the diagnosis information acquired by different types of sensors to judge whether the turbo generator has the rotor winding interturn short circuit fault; the method overcomes the shortcomings of the traditional single diagnosis method; and the sensitivity and the reliability of the rotor winding interturn short circuit fault detection of the turbo generator are improved.

The invention provides a method for user identification. The method comprises the following steps of receiving a user login request, wherein the user login request comprises more than one item of characteristic data of a user; computing a characteristic value corresponding to each item of characteristic data according to a preset function corresponding to each item of characteristic data in more than one item of characteristic data; computing a threshold corresponding to each item of characteristic data according to pre-stored sample data, wherein the threshold is used for distinguishing a characteristic value of a normal account and a characteristic value of an abnormal account; distributing a coefficient to the characteristic value corresponding to each item of characteristic data according to the threshold and the characteristic value of each item of characteristic data; and computing a first risk value corresponding to the user according to the characteristic value corresponding to each item of characteristic data, the coefficient distributed by the characteristic value corresponding to each item of characteristic data and a preset weight value corresponding to each item of characteristic data. The invention also provides a device for user identification. According to the method and the device, the risk of the user can be effectively identified, and the accuracy rate is high.

The invention relates to a method and device for identifying the position of news in a format of an electronic report. The method comprises the following steps: eliminating a header area of a layout; carrying out graying and binarization processing to obtain a binarized image, calculating pixel points of the binarized image one by one pixel row, carrying out eigenvalue assignment on the corresponding pixel row of the blank picture, and obtaining an eigenvector according to an eigenvalue so as to determine an area where the headline of the headline is located; determining the ordinate of the position where the headline is located according to the layout and the headline area, and obtaining the coordinate of the headline according to the abscissa and the ordinate; and obtaining a coordinate range of the article hot area on the layout, and if the coordinate is within the coordinate range, determining the article hot area as the position of the headprint news. Through processing, analysis and calculation of the layout, identification is carried out according to the principle that the maximum title is the headprint, so that the method is more in accordance with logic in principle, and the news identification of the headprint of the electronic report is more accurate, and the maintenance cost is low.

A method, computer program product, and computer system for identifying, by a computing device, a plurality of sources. One or more feature values of a plurality of features may be assigned to a first source of the plurality of sources. One or more feature values of the plurality of features may be assigned to a second source of the plurality of sources. A first score for the first source and a second score for the second source may be determined based upon, at least in part, the one or more feature values assigned to the first source and the second source. One of the first source and the second source may be selected for spatial processing based upon, at least in part, the first score for the first source and the second score for the second source.

The invention discloses a method and device for storing and retrieving feature values. The method includes: extracting feature values ​​from images to obtain feature values; judging whether the file is full for a file to be written with feature values, and if not , write the feature value into the file; if yes, create a new file, write the feature value into the newly created file; assign a search identifier to the feature value according to the order in which the feature value is written. Since in the embodiment of the present invention, after the feature value corresponding to the image is determined, the feature value is written into the file, and the feature value is assigned a search identifier according to the order in which the feature value is written, and when the feature value is retrieved later, it can , input the search identification range information to retrieve the corresponding file, the feature value in the file is the retrieved feature value, and realize the batch search of the feature value. Therefore, it is suitable for batch storage and retrieval of dynamic library scenarios and static library scenarios.

A large-scale image library retrieval method based on the optimal K-means hash algorithm, comprising: selecting some images from the image library to be retrieved as a training image set, first extracting the GIST features of the images in the training set; Preprocessing of feature value assignment; dividing the preprocessed feature data into multiple subspaces; training the codebook and codebook encoding of the subspace for each subspace; processing the feature data in the image library to be retrieved and the query image Corresponding to the training process, extract the GIST features of the retrieval image and the query image respectively, then calculate the hash codes of the features of the image to be retrieved and the query image, and then calculate the Hamming distance between the feature codes of the image to be retrieved and the feature code of the query image , so as to quickly retrieve similar images. The invention has good universality, not only reduces the storage space of data but also improves the retrieval efficiency of query.

The embodiment of the invention discloses a task data processing method and device, a readable storage medium and electronic equipment. The method comprises the following steps: determining a currentpressure attribute of a target object and a matching characteristic value of each piece of task information corresponding to the target object, determining a target diffusion threshold according to the current pressure attribute, and allocating each piece of task information according to the target diffusion threshold and the matching characteristic value of each piece of task information. According to the embodiment of the invention, the target diffusion threshold for task classification can be dynamically determined according to the pressure condition of the current target object, so that the classified task information is distributed in different distribution modes, the task receiving rate of the task information corresponding to the target object is ensured, and the task processing efficiency of a system is further improved.

The present invention relates to a positioning system and method using a combination of results of multimodal sensor-based location recognition. The system according to an embodiment comprises: a sensor unit having a multimodal sensor formed therein to sense obstacle information about an obstacle located around an object; a database unit having a multimodal database built therein, wherein multiple pieces of location-determining information to locate the object are stored in the database; a feature-detecting unit which detects a feature value by performing a feature point-based matching operation of finding a feature point in the obstacle information to recognize an obstacle, and performing a model-based matching operation of comparing a feature point with the multiple pieces of location-determining information to recognize an obstacle having the matching feature point; a fuzzy-mapping unit which corrects the feature value by applying a weight value to the feature value, calculates a value for an area where the object is located and a value for the location of the obstacle according to the corrected feature value, and applies fuzzy logic, thereby generating an initial map into which the area where the object is located and the location of the obstacle are to be incorporated; and a positioning unit for locating the object by using the map into which the area where the object is located and the location of the object have been incorporated.

The present application provides an image processing method and device, wherein the method includes: extracting X magnitude images of the image to be processed, and fusing multiple magnitude images in different directions on the same scale to obtain Y fused Amplitude image; encode the Y fused amplitude images, assign weights to each feature value generated after encoding according to the local characteristics of each amplitude image, and obtain the Y fused amplitude images corresponding to Y statistical histograms with weight information, cascading the Y statistical histograms to form a feature vector of the image to be processed. The technical solution provided by the application is applicable to image recognition in a complex environment, and realizes a more universal image recognition mechanism.

The invention discloses a real-time super-resolution method and system based on an FPGA, and relates to the technical field of image processing. The method comprises the steps of splitting each frameof image in a medium with a relatively low resolution into sub-images for performing super-resolution processing; allocating a processing module according to an eigenvalue of the current frame of thesub-image; if the eigenvalue is relatively high, performing calculation by a neural network module; and if the eigenvalue is relatively low, performing calculation by an interpolation module. An output buffer outputs the high-resolution sub-images obtained by calculation and recombines the sub-images into a high-resolution image, so that a user can see the real-time ultra-high-definition image.