366 results about "A posteriori probability" patented technology

A method is disclosed for classifying an input pattern into an associated class through use of a compound classifier. Data pertaining to preselected features present within the input pattern are extracted. A discriminant value for each of a plurality of classes is then determined via a first classification technique. This value reflects the relative likelihood that a class is the associated class. The class with the highest relative likelihood is selected. A confidence value is generated via a second classification technique. This confidence value is reflective of the a posteriori probability that the selected class is the associated class. The selected class is rejected if the determined confidence value is below a predetermined threshold value.

The invention discloses a method for splitting a news video program and a method and a system for cataloging news videos. The method for splitting the news video program comprises the following steps of: sequencing detection results according to a time sequence to obtain an event sequence by detecting characteristic information of titles of the news video, a headline, characteristic information of comperes, lens transformation, a mute point of an audio, a switching point, a keynote period sudden change point and the like; briefing the event sequence by adopting a preset symbol set and a production rule, and judging rough positions of start points and end points of news sections in the event sequence; calculating a union posterior probability of a start position of each news section near the rough start position according to the event sequence, selecting the moment with the maximum posterior probability as the accurate start position of each news section, splitting the news video, and thus obtaining the news video sections. According to the method, the adopted algorithm is stable and effective; the structural information in the news video can be summarized effectively; the accurate positions of splitting points of the news sections can be determined; and the news video can be split stably and accurately.

A computerized method for automatic document classification based on a combined use of the projection and the distance of the differential document vectors to the differential latent semantics index (DLSI) spaces. The method includes the setting up of a DLSI space-based classifier to be stored in computer storage and the use of such classifier by a computer to evaluate the possibility of a document belonging to a given cluster using a posteriori probability function and to classify the document in the cluster. The classifier is effective in operating on very large numbers of documents such as with document retrieval systems over a distributed computer network.

A maximum a posteriori probability (MAP) block decoder for decoding a received data block of input samples. The MAP block decoder segments the received data block into at least a first segment and a second segment and calculates and stores alpha values during forward processing of the first segment. The MAP block decoder uses a first selected alpha value calculated during forward processing of the first segment as initial state information during forward processing of the second segment. The first and second segments may overlap each other, such that the last M samples of the first segment are the same as the first M samples of the second segment.

An approach is provided for processing structure Low Density Parity Check (LDPC) codes. Memory storing edge information and a posteriori probability information associated with a structured parity check matrix used to generate Low Density Parity Check (LDPC) coded signal are accessed. The edge information represent relationship between bit nodes and check nodes, and are stored according to a predetermined scheme that permits concurrent retrieval of a set of the edge information.

A probabilistic framework for acoustic-phonetic automatic speech recognition organizes a set of phonetic features into a hierarchy consisting of a broad manner feature sub-hierarchy and a fine phonetic feature sub-hierarchy. Each phonetic feature of said hierarchy corresponds to a set of acoustic correlates and each broad manner feature of said broad manner feature sub-hierarchy is further associated with a corresponding set of acoustic landmarks. A pattern recognizer is trained from a knowledge base of phonetic features and corresponding acoustic correlates. Acoustic correlates are extracted from a speech signal and are presented to the pattern recognizer. Acoustic landmarks are identified and located from broad manner classes classified by the pattern recognizer. Fine phonetic features are determined by the pattern recognizer at and around the acoustic landmarks. The determination of fine phonetic features may be constrained by a pronunciation model. The most probable feature bundles corresponding to words and sentences are those that maximize the joint a posteriori probability of the fine phonetic features and corresponding acoustic landmarks. When the hierarchy is organized as a binary tree, binary classifiers such as Support Vector Machines can be used in the pattern classifier and the outputs thereof can be converted probability measures which, in turn may be used in the computation of the aforementioned joint probability of fine phonetic features and corresponding landmarks.

In a Maximum A posteriori Probability decoding (MAP decoding), a correlation and a deviation of noises for past and future states which depend on input signal patterns in past N bits and future Q bits are calculated by training by a noise correlation arithmetic operating unit 84 and they are stored. Upon reproduction, in a white noise arithmetic operating unit 91, white noise values for the past and future states in which colored noises are converted into white noises are obtained by using the stored correlation and deviation of the noises. In an input signal arithmetic operating unit 92, an input signal (channel information) Λc(y_k|S^m_k) of the MAP decoding is calculated from the white noise values and the deviation for the past and future states. A likelihood in the MAP decoding is obtained from the input signal.

A probabilistic digital signal processor for medical function is described. Initial probability distribution functions are input to a dynamic state-space model, which operates on state and/or model probability distribution functions to generate a prior probability distribution function, which is input to a probabilistic updater. The probabilistic updater integrates sensor data with the prior to generate a posterior probability distribution function passed to a probabilistic sampler, which estimates one or more parameters using the posterior, which is output or re-sampled in an iterative algorithm. For example, the probabilistic processor operates using a physical model on data from a medical meter, where the medical meter uses a first physical parameter, such as blood oxygen saturation levels from a pulse oximeter, to generate a second physical parameter not output by the medical meter, such as a heart stroke volume, a cardiac output flow rate, and/or a blood pressure.

The invention provides a countercheck method for automatically identifying a speaker aiming to voice deception, which is a voice anti-spoofing technology based on a method combining various features and a plurality of sub-systems. According to the invention, the serial features of the posterior probability of a phoneme in the phonological level and the MFCC features of voice level or MFDCC features of phase level are combined, thus the performance of the system is significantly enhanced. By combining the provided i-vector sub-system and OpenSMILE (open Speech and Music Interpretation by Large Space Extraction criterion containing voice and rhythmic information, the final presentation of the system is further enhanced. To a back-end model, the development datum are used; and under the situation of knowing deceptive attacks, a two-level support vector machine has better performance compared with one-level cosine similarity or PLDA evaluations, while the one-level evaluation approach has better robustness under the situation without seeing the test datum and knowing the deceptive conditions.

A method according to the invention separates multiple audio signals recorded as a mixed signal via a single channel. The mixed signal is A/D converted and sampled. A sliding window is applied to the samples to obtain frames. The logarithms of the power spectra of the frames are determined. From the spectra, the a posteriori probabilities of pairs of spectra are determined. The probabilities are used to obtain Fourier spectra for each individual signal in each frame. The invention provides a minimum-mean-squared error metho or a soft mask method for making this determination. The Fourier spectra are inverted to obtain corresponding signals, which are concatenated to recover the individual signals.

A parameter estimation method for estimating a parameter by estimating maximum a posteriori probability for input data. Operation for the input data is expressed by an inner product for the input data and the inner product is replaced by a Kernel function. By using the calculation result of the kernel function, a parameter is estimated. The method includes a step (offline operation) for learning a correlation between a plurality of learning input data in which a parameter to be estimated is known and the parameter corresponding to each of the learning input data; and a step (online operation) for estimating the parameter for the estimating input data in which a parameter to be estimated is unknown, by using the learned correlation.

The invention relates a self-adaptive semi-supervised network traffic classification method, system and equipment. The method comprises the following steps: acquiring a network stream, extracting thestream feature with the preset fixed quantity in each network stream to obtain a network stream feature vector; computing the centroid of the network stream feature vector in each type according to the marked network stream feature vector, thereby obtaining a vector set M; performing self-adaptive semi-supervised k-means clustering by taking the vector set M as an initial center point; mapping theobtained network stream in each type of cluster to the belonged traffic type according to the maximum posterior probability; taking the traffic cluster of the known type as the training data to trainan online traffic classifier. The invention further relates to a system, the system comprises an acquisition module, a vector set processing module, a clustering module, a classification module, andan output module. The invention relates to the equipment. The equipment comprises a processor, a memorizer, and a computer program stored on the memorizer and capable of being run on the memorizer.

Methods for processing overhead imagery of a vessel include the step of determining an initial classification and classification probability based on the vessel length and length-to-width ratio. Next, mutually exclusive deck features can be extracted from the image. For several embodiments, the extracted deck features that can be spherical tanks, hatches and containers that are stored on deck. The initial classification probability is then weighted using the results of the deck feature extraction step to yield a posterior classification probability for the ship image. If the posterior classification probability is above a predetermined value, the image is assigned a posterior classification. If the posterior probability is below the predetermined value, the vessel image is classified as unknown, and the gross tonnage of the vessel is calculated using the length and width of the vessel.

A CDMA radiocommunication signals receiver for receiving signals obtained from spectrum symbols spread using pseudo-random sequences and having been propagated along a number of paths. The receiver includes a filter configured to restore L unspread signals for each symbol, corresponding to L different paths, a calculating circuit configured to calculate L estimates of the L different paths, and a demodulator configured to process each of the L unspread signals using the corresponding L estimates to obtain L path contributions. Also included is an adder configured to form a sum of the L path contributions and for outputting an estimate of a received symbol, and a decision circuit configured to make a decision about a value of the received symbol based on a value of the estimate of the received symbol output by the adder. Further, the receiver processes blocks of N symbols, each block having data symbols and control symbols, each symbol being identified by a rank k that it occupies in the block, where k varies from 0 to N-1. Also, for each path identified by an index l, where l varies from 0 to L-1, and for each block, the receiver considers a vector Cl with N components that characterizes the path during the block, and the receiver defines a vector base BK, vectors of the vector base BK being N eigenvectors of the matrix E [ClCl<.T>], each vector Cl being decomposed in the vector base, where decomposition coefficients denoted GlK form independent random Gaussian variables. In addition, coefficients GlK, define a vector Gl with N components for each path l, and the calculating circuit estimates each vector Gl, using an iterative process based on EM estimation-maximization algorithm based on a maximum a posteriori probability criterion.

The invention discloses a laminated and partitioned irregular low density parity check (LDPC) code decoder and a decoding method in the technical field of communication. An external information storage unit outputs a soft value transmitted to an information node by a last iterated check node to a decoding processing module. A cyclic shift register transmits a posterior probability likelihood ratio update value of the information node to the decoding processing module. The decoding processing module transmits the check update value in the iteration to the external information storage unit, andsimultaneously transmits the posterior probability likelihood ratio update value of the information node to the cyclic shift register through a decoding processing module interweaving network. The decoder is suitable for decoding all quality control (QC) LDPC codes, and all the partitioned LDPC code words support decoding; the decoder has no stream competition conflict, and has better throughput performance and relatively simple working time sequence; and the consumption of the interweaving network of huge resources is not needed, many hardware resources are saved, and the resource consumption of the whole decoder is relatively low. The decoding supporting parallelism degree can be flexibly changed.

A communication system having a program of machine-readable instructions for solving an ILS problem, tangibly embodied on a computer readable memory and executable by a digital data processor, to perform actions directed toward outputting a set of a-posteriori probability vectors.

Multilevel soft-equalizer detectors, such as a maximum a posteriori probability (MAP) detector, suitable for use in polarization multiplexed optical communications using multilevel modulations and coherent detection are disclosed. Detection systems and methods may consider two symbols transmitted over two orthogonal polarization states as a two-component symbol, which is effective in eliminating the bit error ratio (BER) floor phenomenon introduced by conventional soft equalizers.

The invention provides a probabilistic model based automatic target tracking method for a moving camera, which comprises the following specific steps of: firstly, initializing start tracking parameters and real-time moving target tracking parameters; secondly, utilizing a speed difference function to position an initial moving area, and adopting an improved optimal clustering function to segment a moving target; thirdly, establishing a color probability histogram model for the moving target detected by the current frame and a background search window, and utilizing a posteriori probability function and a color histogram similarity function to adjust the size and position of the target; and finally, using a continuously adaptive mean shift method to find the next-frame moving target to achieve the real-time tracking of the moving target in the dynamic background. In the invention, the reliability of moving target tracking is improved, and the problems of selection of initial targets and multi-target tracking, and the problems that the target and background have the same color and the target color changes can be effectively solved under the moving camera.

A system and method for noise reduction applied to a speech recognition front-end. An output of a front-end is optimized by giving, as a weight to the output for each band, a confidence index representing the remarkableness of the harmonic structure of observation speech. In a first method, when clean speech is estimated by executing MMSE estimation on a model that gives a probability distribution of noise-removed speech generated from observation speech, the posterior probability of the MMSE estimation is weighted using the confidence index as a weight. In a second method, linear interpolation is executed, for each band, between an observed value of observation speech and an estimated value of clean speech, with the confidence index serving as a weight. The first method and the second method can be combined.

The invention discloses a SAR image change detection method based on a priori, a fusion gray level and a textural feature. By using the method of the invention, problems that a Gaussian model can not completely fit distribution of a difference graph and change detection accuracy is low because only pixel gray level information of the SAR image is used are mainly solved. The method comprises the following realization steps that (1) two time phase SAR images which are registered and corrected are read in; (2) a wavelet fusion strategy is performed on the two images so as to construct the difference graph; (3) a classified priori probability of the difference graph is calculated; (4) the gray level of the difference graph and the texture information are fused so as to acquire an observed quantity likelihood probability; (5) the classified priori probability and the observed quantity likelihood probability are used to calculate a posteriori probability; (6) a maximum posteriori probability criterion is used to divide the difference graph into a change type and a non-change type; (7) a step (3) to a step (6) are repeated till a terminal condition is satisfied and a final change detection result is output. The method of the invention has the advantage that change detection precision to the SAR image is high. The method can be used to extract and acquire change detail information of the SAR image.

The invention provides an AUV (Autonomous Underwater Vehicle) multi-source information fusion method andequipment based on a factor graph, which are used for realizing navigation and positioning of anAUV integrated navigation system. According to the method, an information fusion problem of the AUV integrated navigation system is converted into a maximum posteriori probability problem for solvingjoint probability distribution of multivariate random variables by utilizing a factor graph theory, and the maximum posteriori probability problem is expressed by using a factor graph model. Optimalestimation of the navigation state is achieved according to transmission and updating of a factor graph message transmission algorithm, namely a sum product algorithm, in a factor graph model of the AUV multi-source information integrated navigation system. The method effectively solves the problems of asynchronization and short-time failure of updating frequencies of different navigation information sources in an AUV integrated navigation system, and is suitable for underwater complex and changeable environments and task requirements.

The invention discloses an SAR image segmentation method on the basis of the HMT model in the Shearlet domain, which pertains to the technical field of image processing and mainly aims at solving the problem that the application of the traditional multi-scale geometrical analysis in SAR image segmentation is easy to result in poor regional uniformity and disorder edges. The segmentation process comprises the steps of extracting feature areas (I0, I1, and the like, and IC) in the SAR image to be segmented, calculating the Shearlet transformation coefficients (S0, S1, and the like, and SC) of the feature areas, utilizing the EM algorithm to obtain the HMT model parameter set (Theta1, Theta2, and the like, and ThetaC) in the Shearlet domain of various feature areas, carrying out Shearlet transformation to the SAR image to be segmented to obtain an image coefficient S, utilizing the feature coefficients (S0, S1, and the like, and SC) to calculate likelihood values (Lhood, Lhood, and the like, and Lhood) corresponding to the SAR image coefficient S in each scale, calculating initial segmentation results (MLseg, MLseg, and the like, and MLseg) of the likelihood values in each scale according to the maximum likelihood rule, carrying out fusion to the initial segmentation results by maximizing a posteriori probability criterion and taking the fused image of the scale at the first level as a final segmentation result. The method has the advantages of high convergence rate, good regional uniformity of segmentation result and completely retained information, and can be applied to SAR image target identification.

A decoding process for decoding a received block of N systematic binary data samples or N systematic duobinary data samples using a maximum a posteriori probability (MAP) decoding algorithm. The decoding process calculates a set of four log-likelihood values using the corresponding forward state metric, reverse state metric, and branch metric values for each of N/2 pairs of systematic binary data or each of N/2 pairs of duobinary data in the received block. The decoding process also calculates, for each set of four log-likelihood values a delta value corresponding to the difference between the largest and the second largest of the four log-likelihood values in each set. The decoding process repeats for at least a second iteration. The decoding process is stopped based on a plurality of delta values calculated during two consecutive iterations.