39 results about "Entropy rate" patented technology

The invention discloses a fault identification method of an analog circuit based on an improved extreme learning machine, wherein the method comprises steps: the voltage eigenvectors of analog circuits at different corner frequencies are used as inputs, and then input vectors corresponding to each hidden layer neuron are selected based on each entropy rate, and weights and offsets that reach the highest correlation degree are generated through the multi-dimensional particle swarm optimization algorithm, then the appropriate parameters are found by PSO algorithm and iterative operation, so as to construct an efficient hidden layer model; finally, the hidden layer model is trained to identify the fault of analog circuit, and the fault of analog circuit can be identified. The model has the characteristics of high fault identification accuracy and high speed.

The invention discloses a moving vehicle detection method including a pause state in an aerial photograph video of an unmanned aerial vehicle. At first, multilevel preprocessor is used to preprocess that image, as the image is filtered and denoised, the aim of this paper is to improve the characteristics of uneven illumination in aerial video of unmanned aerial vehicles, the entropy rate superpixel segmentation algorithm with superpixel regularity is then used to extract the superpixels from the image, dictionary learning algorithm based on correlation and double sparsity obtains the color ofsuper pixel, texture, geometry, a fusion similarity matrix is constructed for image segmentation, using the texture and geometric features of the RAG fusion region of successive multi-frame images toperform image matching, the motion information of the RAG fusion region is obtained. Furthermore, an image foreground and background segmentation method based on a motion similarity map are proposed to realize the detection of a moving vehicle with a pause state, thereby greatly reducing the difficulty of detecting the moving vehicle. The algorithm of the invention is simple and quick, and can effectively detect the moving vehicle with a pause state in a video shot by an unmanned aerial vehicle.

A method, system and computer program product are provided for progressively encoding a digitized color image with M distinct colors allocated to a set of pixels using a data processing system, the digitized color image being provided by assigning each of the M distinct colors to at least one pixel in the set of pixels. This involves: (a) initializing a tree structure by providing at least one starting leaf node comprising a subset of pixels from the set of pixels; (b) determining at least one representative color for each starting leaf node; and (c) growing the tree structure by (i) selecting a leaf node n to become a non-leaf node n linked to two new leaf nodes based on an associated achievable cost, wherein the associated achievable cost is determined by 1) determining an associated change in distortion resulting from turning the leaf node into the non-leaf node linked to the two new leaf nodes; 2) determining an associated increase in entropy rate resulting from turning the leaf node into the non-leaf node linked to the two new leaf nodes; and, 3) determining the associated achievable cost based on the associated change in distortion and the associated increase in entropy rate; (ii) creating the two new leaf nodes by allocating each pixel in n to one of the two new leaf nodes; (iii) determining a representative color for each of the two new leaf nodes; and (iv) encoding the index information associated with the leaf node n, the representative colors of the two new leaf nodes, and information regarding each pixel in n.

An image is segmented into superpixels by constructing a graph with vertices connected by edges, wherein each vertex corresponds to a pixel in the image, and each edge is associated with a weight indicating a similarity of the corresponding pixels, A subset of edges in the graph are selected to segment the graph into subgraphs, wherein the selecting maximizes an objective function based on an entropy rate and a balancing term. The edges with maximum gains are added to the graph until a number of subgraphs is equal to some threshold.

A system and method for high-speed transfer of small data sets, that provides near-instantaneous bit-level lossless compression, that is ideal for communications environments that cannot tolerate even small amounts of data corruption, have very low latency tolerance, where data has a low entropy rate, and where every bit costs the user bandwidth, power, or time so that deflation is worthwhile. Where some loss of data can be tolerated, the system and method can be configured for use as lossy compression.

The invention discloses a visual attention computing method and a system based on bit entropy rate. The method includes that: a sparse codebook is learned to obtain a sparse codebook primary function; the sparse codebook primary function is adopted to filter the image or video data and obtain a plurality of sub-band feature graphs; a corresponding fully-connected graph is established for each sub-band feature graph; the information transfer method of each fully-connected graph adopts the random walk method, during the random walk process, the significance measurement is performed according to the bit entropy rate to accordingly obtain a bit entropy rate graph corresponding to each sub-band feature graph; the bit entropy rate graphs are added together to obtain the significance graph of the image or video data. As indicated in large amount of experiments, and compared with other methods in prior art, the analysis result of image significance analysis or video significance analysis obtained through the invention is more accurate and is supported by the basis of physiology and psychology.

The invention discloses a hyperspectral image noise label detection method based on super-pixel weight density. The method comprises the following steps: firstly, introducing an entropy rate superpixel (ERS) algorithm into a step of obtaining context data information for training samples (with noise tags); secondly, obtaining Euclidean distance measurement through a super-pixel region; next, introducing a Gaussian weighting algorithm into Euclidean distance calculation, defining a weighting coefficient of distance information of each class, finally, adaptively detecting a training sample through a clustering algorithm based on a density peak value, and evaluating the effectiveness of the proposed detection method by utilizing a support vector machine. Experiments carried out on several real hyperspectral data sets show that the provided label detection method can effectively improve the performance of a classifier with label noise pollution in an initial training sample, and has very great theoretical significance and practical application value.

The invention discloses an automatic random-walk CT lung parenchyma image segmentation method based on prior information. The method comprises the following steps that an original CT image is input, and binary discrete wavelet transformation is carried out on the input CT image to obtain an image M; a chest area is extracted from the image M in an entropy based superpixel method, and then processed to obtain an image T; the amount and position of seed points of the image T are obtained by utilizing the prior information guided by anatomy knowledge, and random-walk segmentation is carried out on the obtained seed points to obtain an initial parenchyma contour of the lung; and local mis-segmented areas of the initial parenchyma contour of the lung are repaired in a curvature rectification algorithm. Thus, the method of the invention can be used to solve the technical problem that in the present random-walk technology, the segmented lung parenchyma image is greatly influenced by the amount and positions of the seed points when the seed points are selected by manual interaction.

The invention relates to a large-format remote sensing image building extraction method for mixed scenes. It includes the following steps: step 1, preprocessing the remote sensing image; step 2, performing superpixel segmentation based on entropy rate; step 3, extracting three features of shadow, homogeneity and strong edge; step 4, generating shadow layer, Homogeneity layer and strong edge layer; step 5, extract buildings on different layers; step 6, building overlay. The beneficial effect of the present invention is that more features and deeper semantic information can be extracted, which can be adapted to extract buildings in a larger geographical area and be used to realize related services such as update of urban basic geographic databases.

The invention belongs to the technical fields of information security and true random number generators, and specifically relates to a method for testing the maximum entropy rate of true random number generators. The method for testing the maximum entropy rate of a true random number generator disclosed by the invention includes four steps: parameter selection, data collection, block calculation, and analysis and fitting. The present invention utilizes the characteristic that the average entropy rate calculated after the random data flow is divided into blocks is saturated with the decrease of the data block length and the increase of the data rate, and performs linear fitting on the relationship curve between the saturated entropy rate and the data block length, and simulates The corresponding function value of the straight line when the data block length is 1 is the maximum entropy rate. The invention can effectively solve the problem that the maximum entropy rate cannot be uniquely determined in the test of the true random number generator.

The invention relates to a method for testing and diagnosing the cause of performance attenuation of a lithium-ion battery, comprising measuring the ohmic impedance and relaxation impedance of a battery to be tested and a reference battery during charging and discharging under a certain state of charge; The temperature entropy coefficient of the battery to be tested and the reference battery; calculate the ohmic impedance growth rate and relaxation impedance growth rate and the absolute value of the temperature entropy coefficient growth rate of the battery relative to the reference battery in the charging and discharging process; Comparing the ohmic impedance growth rate and relaxation impedance growth rate and the absolute value of the temperature entropy coefficient growth rate of the battery under test relative to the reference battery, the cause of battery performance attenuation can be judged. The invention avoids the influence of other factors introduced into the assembly process after the battery is disassembled, and greatly improves the accuracy of testing and judgment; and the operation process and method are simple, which is convenient for testing multiple battery samples and establishing a corresponding database, which can reduce battery performance attenuation. Analysis provides reliable reference data.

The invention discloses a visual attention computing method and a system based on bit entropy rate. The method includes that: a sparse codebook is learned to obtain a sparse codebook primary function; the sparse codebook primary function is adopted to filter the image or video data and obtain a plurality of sub-band feature graphs; a corresponding fully-connected graph is established for each sub-band feature graph; the information transfer method of each fully-connected graph adopts the random walk method, during the random walk process, the significance measurement is performed according tothe bit entropy rate to accordingly obtain a bit entropy rate graph corresponding to each sub-band feature graph; the bit entropy rate graphs are added together to obtain the significance graph of the image or video data. As indicated in large amount of experiments, and compared with other methods in prior art, the analysis result of image significance analysis or video significance analysis obtained through the invention is more accurate and is supported by the basis of physiology and psychology.

The invention discloses a color image tamper detection method based on local quaternion index moment through which algorithm time complexity can be greatly reduced and a replication tamper area can be accurately detected. Self-adaptive super-pixel partitioning is performed on an image to be detected after Gaussian smoothing filtering preprocessing by utilizing a super-pixel segmentation algorithm based on entropy rate with combination of non-subsample Shearlet transform; the quaternion index moment is utilized to express Sifer feature point local area features and similar block matching is performed on all the Sifer feature point local area features of each super-pixel block; a suspected tamper area is determined by utilizing the matching feature points in the similar blocks obtained in the step 2; and the replication tamper area is marked by utilizing morphological filtering operation.