Radar target parameter estimation method based on AIC (automatic information center) compression information acquisition and FBMP (fast Bayesian matching pursuit)

The present invention discloses a multi-parameter multi-object chaotic particle swarm parameter optimization method. The method comprises the steps of (1) determining a target function and a parameter to be optimized, (2) initializing an algorithm, (3) calculating the target function corresponding to each individual in a population, (4) updating an individual history optimal solution, (5) updating a particle velocity and position, (6) updating a global optimal solution set, (7) updating a global optimal solution, and (8) carrying out result judgment. Compared with a common random initialization method and an existing chaotic logistic mapping particle swarm initialization method, according to the method of the invention, the performance of global optimization is improved and the stability is good, compared with a common multi-object weighted optimization method, the Pareto optimal solution technology is employed by the method, and the problem of difficult weight selection in the multi-object method is solved.

Multi-parameter multi-object chaotic particle swarm parameter optimization method

The invention discloses an SAS image segmentation method and system based on an SVM classifier. The method comprises the steps of (101) pre-processing an SAS image to be segmented to acquire grey level histograms and grey level co-occurrence matrices of different areas of the SAS image; (102) extracting a plurality of statistical property parameters from the grey level histograms and a plurality of textural feature parameters from the grey level co-occurrence matrices; (103) inputting the statistical property parameters and the textural feature parameters into the SVM classifier to carry out segmentation on the SAS image to be segmented, wherein the SVM classifier is obtained through training by taking a plurality of statistical property parameters extracted from a grey level histogram of a certain area of the training image and a plurality of textural feature parameters extracted from a grey level co-occurrence matrix of the certain area of the training image as training features; (104) optimizing segmented images output from step (103) by means of mathematical morphology operators so that cavities and isolated points in the segmented images can be removed.

SAS image segmentation method and system based on SVM classifier

The invention discloses an image super-resolution reconstruction method based on cognitive regularization parameters. A human eye cognition theory is used for calculating appropriate regularization parameters for all image blocks, and the regularization parameters are called as the cognitive regularization parameters and introduced into the process of solving sparse representation, then in combination with the super-resolution reconstruction method based on sparse representation, all the blocks are reconstructed, and finally all the reconstructed blocks are recombined to obtain a high-resolution image. The image super-resolution reconstruction method is used for testing in a facial image, and the result shows that the method has a better image reconstruction effect than that of a method using fixed regularization parameters.

Image super-resolution reconstruction method based on cognitive regularization parameters

According to some aspects, a method of searching for content in response to a user voice query is provided. The method may comprise receiving the user voice query, performing speech recognition to generate N best speech recognition results comprising a first speech recognition result, performing a supervised search of at least one content repository to identify one or more supervised search results using one or more classifiers that classify the first speech recognition result into at least one class that identifies previously classified content in the at least one content repository, performing an unsupervised search of the at least one content repository to identify one or more unsupervised search results, wherein performing the unsupervised search comprises performing a word search of the at least one content repository, and generating combined results from among the one or more supervised search results and the one or more unsupervised search results.

Systems and methods for speech-based searching of content repositories

Self-adapting image compressive sampling method based on multi-dimension saliency map

The present invention discloses a texture image classifying method based on the SWBCT, and belongs to the technical field of image processing. The aim is to overcome the deficiency of the validity of the classifying characteristic. The present invention is put forward and used in the classification of the texture image aiming at a contour wave and the problem which is encountered in the texture analysis using the WBCT. The methods of energy, Hu matrix and cooccurrence matrix are respectively used in the taking phase of a child-band characteristic to obtain a better classifying effect. The concrete implementing steps of the invention are as follows: 1, a sample image is inputted and normalized. 2, the normalized sample image proceeds an image decomposition-SWBCT decomposition with multi directions and multi resolutions, and a decomposition child-band is obtained. 3, the Characteristic taking proceeds to the obtained child-band and various characteristics are combined. 4, a classifying method is chosen from the existing classifying methods, and a texture image analysis proceeds for the combined characteristic. The method can be used in mode identifications such as image classification, script identification and sound identification and classification.

Texture image classification method based on SWBCT

The invention discloses a multiscale SAR image segmentation method based on semi-supervised learning, belonging to the technical field of image processing and mainly overcoming the disadvantages of low segmentation accuracy and relatively long operation time of the traditional segmentation methods. The implementation steps are as follows: (1) three-layer wavelet transform and three-layer Contourlet transform are respectively carried out on the images to be segmented to finish image decomposition and a coarse decomposition subband, a sub-coarse decomposition subband and a fine decomposition subband are obtained by merge operation; (2) with respect to the coarse decomposition subband, the method of semi-supervised learning is adopted to finish initial segmentation and obtain the results ofinitial segmentation; and (3) multiscale secondary segmentation based on unsupervised learning is carried out on the results of initial segmentation, the sub-coarse decomposition subband and the finedecomposition subband obtained in step (1) to obtain the final segmentation result. The method improves the accuracy of the segmented images, reduces the misclassification rate and can be used for texture image segmentation, natural image segmentation and medical image segmentation.

Multiscale SAR image segmentation method based on semi-supervised learning

The invention discloses a V-BLAST system detection method based on discrete particle swarm algorithm, which belongs to the technical filed of artificial intelligence and aims to overcome the disadvantages in the prior art, such as high calculation complexity and code error rate. The inventive method can provide better performance of code error rate and reduce the actual complexity. The method is implemented by the steps: 1. initializing a particle swarm with a size of m, randomly generating a position vector of each particle, the dimensional number of the position vector being the same as thetransmitting antenna, wherein the setup variation rate is m, the evolution algebra t is zero and the condition of ending the determined algorithm is that the maximal iterative algebra is ga; 2. calculating the particle adaptability, judging whether the particle adaptability meets the condition of iterative termination, if meeting the condition, the algorithm is ended, otherwise going to the step 3; step 3, updating the particle position, executing the variation operation, updating the individual extreme value and the overall extreme value, and going to the step 2. The invention can be used tosolve the detection problem of vertical layered space-time system in the field of wireless communication.

Discrete particle cluster algorithm based V-BLAST system detecting method

The invention discloses a radar two-dimensional image classifying method based on multi-dimension geometric analysis, belongs to the technical field of image processing, and mainly overcomes the defect that an existing method can not effectively express a radar two-dimensional image. The method comprises the following steps: firstly, a sample image set is input, and each image in the sample imageset is normalized; secondly; three-layer wavelet decomposition and Contourlet decomposition are carried out on each normalized sample image to obtain ten wavelet decomposition subbands and thirteen Contourlet decomposition subbands which are respectively corresponding to the sample image; thirdly, energy characteristics are carried out on the obtained decomposition subbands and are merged by utilizing a characteristic merging method; and fourthly, the merged characteristics are classified by selecting support vector machine (SVM) algorithm. The invention has better classification accuracy rate and lower complexity, and can be used for classification of radar two-dimensional images and texture images as well as identification of bridge targets in SAR images.

Fan Liu

Papers

Self-adapting image compressive sampling method based on multi-dimension saliency map

Texture image classification method based on SWBCT

Multiscale SAR image segmentation method based on semi-supervised learning

Discrete particle cluster algorithm based V-BLAST system detecting method

Method for sorting radar two-dimension image based on multi-dimension geometric analysis