1508 results about "Improved algorithm" patented technology

Disclosed is an improved algorithm for estimating the 3D shape of a 3-dimensional object, such as a human face, based on information retrieved from a single photograph by recovering parameters of a 3-dimensional model and methods and systems using the same. Beside the pixel intensity, the invention uses various image features in a multi-features fitting algorithm (MFF) that has a wider radius of convergence and a higher level of precision and provides thereby better results.

An improved sectored cache replacement algorithm is implemented via a method and computer program product. The method and computer program product select a cache sector among a plurality of cache sectors for replacement in a computer system. The method may comprise selecting a cache sector to be replaced that is not the most recently used and that has the least amount of modified data. In the case in which there is a tie among cache sectors, the sector to be replaced may be the sector among such cache sectors with the least amount of valid data. In the case in which there is still a tie among cache sectors, the sector to be replaced may be randomly selected among such cache sectors. Unlike conventional sectored cache replacement algorithms, the improved algorithm implemented by the method and computer program product accounts for both hit rate and bus utilization.

A physical design automation system produces an optimal placement of microelectronic components or cells on an integrate circuit chip. An initial population of possible cell placements is generated, and repeatedly altered using simulated on or other fitness improvement algorithm to progressively increase the fitnesses (decrease the costs) of the placements. After each alteration step, the fitnesses of the placements are calculated, and less fit placements are discarded in favor of more fit placements. After a termination criterion is reached, the placement having the highest fitness is designated as the optimal placement. Two or more fitness improvement algorithms are available, and are optimally switched from one to the other in accordance with an optimization criterion to maximize convergence of the placements toward the optimal configuration.

Improved algorithms are provided for performing path management for a plurality of select data transmission sessions in a multicast network. The algorithms include: constructing a weighted graph which represents topology of the network; assigning load correlated costs to network links; and computing least cost paths for each of the data transmission sessions which accounts for global bandwidth utilization.

The invention discloses a method for calculating angle of inclination and correcting inclination by using image contour information and combing Hough conversion. The method comprises: a) using an image collection device to obtain a certificate image; b) reducing the obtained certificate image; c) carrying out grey treatment on the certificate image; d) using an Sobel operator module to detect the edges of the gray image to obtain a thicker wide profile image; e) carrying out detailing operation on the treated image to obtain a profile image with thin edges and one pixel width; f) carrying out Hough conversion on the obtained profile image with thin edges to detect the line parameter of a frame, and using detected parameter to calculate to obtain the angle of inclination of an original certificate; and g) using the obtained angle of inclination to rotate the image. The invention fully utilizes reduced image contour information and uses Hough conversion to calculate the angle of inclination, which greatly reduces points participating in operation, improves algorithm efficiency and ensures high accuracy, strong robustness and high reliability of algorithm detection.

The invention belongs to the field of resource aggregation, and particularly relates to a block chain technology-based data sharing solution, which comprises the following steps of: data model architecture design; business intelligent contract design; authority management and identity authentication design; access control mechanism design; data storage structure design. According to the multi-layer block chain model supporting massive data hierarchical authorization transactions, an offline storage mode is adopted, so that the storage space consumption of the block chain is greatly reduced. Comparison experiments verify that the Ethereum data model has great advantages in the aspects of data storage and access control compared with bit coins. An alliance chain consensus algorithm based ona message queue kafka and a private chain consensus algorithm based on an improved PBFT algorithm are provided.

The invention discloses an automatic-driving intelligent vehicle trajectory tracking control strategy based on deep reinforcement learning. For an intelligent vehicle automatic-driving task, accordingto an action-critics structure in a deterministic policy gradient (DDPG) algorithm, a double-action network is adopted to output a steering wheel angle command and a vehicle speed command. A main reviewer network is designed to guide the updating process of the double-action network, and the stragety specifically comprises the steps of describing an automatic driving task as a Markov decision process: < st,at,Rt, st +1 >; initializing a'double-action 'network in the improved DDPG algorithm by adopting a behavior cloning algorithm; pre-training a'reviewer 'network in the deep reinforcement learning DDPG algorithm; designing a training road containing various driving scenes to carry out reinforcement learning online training; and setting a new road to test the trained deep reinforcement learning (DRL) model. The control strategy is designed by simulating the human driving learning process, and automatic driving of the intelligent vehicle in the simple road environment is achieved.

Improved algorithms are provided for performing path management for a plurality of select data transmission sessions in a multicast network. The algorithms include: constructing a weighted graph which represents topology of the network; assigning load correlated costs to network links; and computing least cost paths for each of the data transmission sessions which accounts for global bandwidth utilization.

The invention discloses a neural network method for diagnosing analog circuit failures which is based on a particle swarm algorithm, and comprises the following steps: imposing an actuating signal to an analog circuit to be tested, measuring an actuating response signal in the testing nodes of the circuit, extracting the candidate signal of failure characteristics by implementing noise elimination and then wavelet packet transformation on the measured actuating response signal, extracting the failure characteristics information by further implementing orthogonal principal component analysis and normalization processing on the candidate signal of failure characteristics, and sending the failure characteristics information as samples to the neural network for implementing classification. The method adopts the particle swarm algorithm instead of a gradient descent method in traditional BP algorithms, thus leading the improved algorithm to be characterized in that the algorithm avoids the local minimum problem and has better generalization performance. The BP neural network method for diagnosing the analog circuit failures which is optimized on the basis of particle swarm can obviously reduce iteration times in the algorithm, improve the precision of network convergence, and improve diagnosis speed and precision.

Disclosed is an improved algorithm for estimating the 3D shape of a 3-dimensional object, such as a human face, based on information retrieved from a single photograph by recovering parameters of a 3-dimensional model and methods and systems using the same. Beside the pixel intensity, the invention uses various image features in a multi-features fitting algorithm (MFF) that has a wider radius of convergence and a higher level of precision and provides thereby better results.

A computer-implemented method, computer program product and data processing system to improve runtime performance of executable program code when executed on the data-processing system. During execution, data is collected and analyzed to identify runtime behavior of the program code. Heuristic models are applied to select region(s) of the program code where application of a performance improvement algorithm is expected to improve runtime performance. Each selected region is recompiled using selected performance improvement algorithm(s) for that region to generate corresponding recompiled region(s), and the program code is modified to replace invocations of each selected region with invocations of the corresponding recompiled region. Alternatively or additionally, the program code may be recompiled to be adapted to characteristics of the execution environment of the data processing system. The process may be carried out in a continuous recursive manner while the program code executes, or may be carried out a finite number of times.

The invention discloses a method for identifying opinion leaders in a social network based on topic link behaviors of users, comprising constructing a social network directed graph, amending edge-weight of the network, for each network node i, calculating the number of shortest paths passing through the node i, calculating the length of the shortest paths, weighting edges of the node, and calculating weighted sum B(i). Opinion leaders based on topic tracking are composed of nodes which have a larger value of B(i). The method of the invention adopts a calculation method of link relations in social network user topics, which reduces scale of network nodes and accordingly decreases system overhead and improves algorithm accuracy and recall rate.

A computer-implemented method, computer program product and data processing system to improve runtime performance of executable program code when executed on the data-processing system. During execution, data is collected and analyzed to identify runtime behavior of the program code. Heuristic models are applied to select region(s) of the program code where application of a performance improvement algorithm is expected to improve runtime performance. Each selected region is recompiled using selected performance improvement algorithm(s) for that region to generate corresponding recompiled region(s), and the program code is modified to replace invocations of each selected region with invocations of the corresponding recompiled region. Alternatively or additionally, the program code may be recompiled to be adapted to characteristics of the execution environment of the data processing system. The process may be carried out in a continuous recursive manner while the program code executes, or may be carried out a finite number of times.

The invention discloses a geomagnetic indoor positioning system and method based on a self-adaptive particle filter algorithm. The geomagnetic indoor positioning system comprises the novel self-adaptive particle filter algorithm and an efficient geomagnetic fingerprint data collecting unit. The efficient geomagnetic fingerprint data collecting unit has the main function of quickly collecting indoor geomagnetic signals by using a mobile phone magnetometer and converting the indoor geomagnetic signals into a geomagnetic fingerprint model to be stored. The novel self-adaptive particle filter algorithm is the improvement to an existing filter algorithm to improve the robustness, the precision and the usability. The key technology of the self-adaptive particle filter improvement algorithm mainly comprises a self-adaptive behavior model, a novel measurement model, a self-adaptive resampling model and a positioning precision estimation and positioning failure detection model. The geomagnetic indoor positioning system is suitable for various smart phones integrated with acceleration sensors, gyroscopes and magnetometers. The geomagnetic indoor positioning system has the advantages that the facing directions, the placing positions and using of the smart phones are not limited, and positioning precision is high.

The present invention relates to the field of target tracking and more generally to a method employing improved algorithms, which achieve excellent tracking performance for a high-g maneuvering target. The two-model Interacting Multiple Model algorithm and the Interacting Acceleration Compensation algorithm will be modified by introducing adaptive factors through the detection of the normalized innovation squared which is chi-square probability distributed. The implementation results show that the modified algorithms can handle the target sudden maneuver better and are more accurate than the original algorithms.

The invention discloses a space-ground integrated network resource allocation method based on an improved genetic algorithm, comprising the following steps: defining parameters and decision variables;establishing a multi-objective constraint model; and allocating resources based on the improved genetic algorithm. The method considers the allocation of multiple resources, so that the resource utilization rate of the space-ground integrated network is significantly improved. The improved selection mechanism effectively retains elite individuals and speeds up the convergence of the improved genetic algorithm. The shortest time for completing all tasks is taken as a objective function, and the priorities of the tasks are considered at the same time, so that the rationality of resource allocation is effectively improved; and the elite retention strategy is combined with the roulette strategy to improve the selection mechanism, adaptive crossover and mutation operators are designed to improve the existing genetic algorithm, and the improved algorithm can effectively avoid the shortcomings of poor local optimization ability of the genetic algorithm and easiness to fall into local optimum, prevent the loss of the optimal solution and effectively improve the optimization speed.

The invention discloses a FastSLAM algorithm based on an improved resampling method and particle selection. The algorithm comprises the following steps: (1) a robot predicts robot poses according to input control data and updates the robot poses and road signs according to measuring data combined with a measuring model of an external sensor of the robot; (2) the robot poses are predicted by the calculation of a particle filter, and a particle resampling criterion is amended according to an effective particle number, particle weighting covariance and particle-measuring residual consistency; (3) new particles are generated by using an index grad method and a crossover operator; (4) the robot is positioned, and a map is created according to the generated new particles. The invention improves the particle resampling criterion and a new particle-generating method in the FastSLAM algorithm, thereby obviously improving the estimation consistency of the FastSLAM algorithm to the robot poses and simultaneously improving the precision of the robot positioning and map creation.

The invention belongs to the technical field of unmanned vehicle navigation, particularly relates to an unmanned vehicle path planning method based on an improved A * algorithm and deep reinforcementlearning. The method aims to give full play to the advantages of global optimization of global path planning and real-time obstacle avoidance of local planning, improve the rapid real-time performanceof an A * algorithm and the complex environment adaptability of a deep reinforcement learning algorithm, and rapidly plan a collision-free optimal path of an unmanned vehicle from a starting point toa target point. The planning method comprises the following steps: establishing an initialized grid cost map according to environmental information; planning a global path by using an improved A * algorithm; designing a sliding window based on the global path and the performance of the laser radar sensor, and taking the information detected by the window as the state input of the network; on thebasis of a deep reinforcement learning method, using an Actor-Critic architecture for designing a local planning network. According to the invention, knowledge and a data method are combined, an optimal path can be obtained through rapid planning, and the unmanned vehicle has higher autonomy.

The invention relates to a method for measuring coal bunker material level based on laser and binocular vision and a device thereof, belonging to the field of machine vision. The device comprises two CCD cameras with the same parameters, two lenses with consistent parameters, one or a plurality of lasers, a light source, a processing element and an explosion-proof case provided with a window. A rotating mechanism is also included when one laser is used. The method includes that the laser emits variable size laser beams, so as to form a light spot on coal surface, image parallax of the light spot between the two cameras is utilized to measure material level, and an image pair is acquired by the two cameras; corrosion expansion pre-treatment is carried out on the obtained images by utilizing an improved algorithm; for the two images after being pretreated, a new rapid stereoscopic vision algorithm is utilized to calculate a depth map; depth is obtained according to the depth map, and finally high efficiency, real time and accurate coal bunker material level measurement is realized. The device has two measuring modes, namely real-time detection mode and sensor detection mode.

The invention discloses a robot scene self-adaptive pose estimation method based on RGB-D camera. scene two-dimensional color image information of adjacent frames and space depth information corresponding to pixels of the two-dimensional color image are acquired based on a RGB-D camera; when the two-dimensional color image feature points are sufficient, ORB operators are adopted to extract features, the matching strategy provided by the invention is adopted to carry out accurate matching, and three-dimensional pose estimation is solved based on the pose estimation algorithm of the matching feature points; when the feature points are insufficient, the improved ICP algorithm provided by the invention is adopted to solve the three-dimensional pose estimation; then, a complete switching criterion is designed to fuse the two pose estimation methods; finally, pose estimation obtained through the two methods is optimized through a light beam adjustment algorithm, and finally smooth and accurate three-dimensional pose estimation is obtained. The three-dimensional pose estimation algorithm has the outstanding advantages of high robustness, high precision, small calculation amount, adaptability to different scenes and the like.

The invention discloses a constraint condition-based random forest recommendation algorithm. The algorithm comprises the steps of generating a binary decision tree according to a data set in a given initial training set; randomly selecting features to form a random forest; measuring time efficiency of a decision tree algorithm; measuring efficiency of a random forest algorithm; and improving the random forest algorithm. Compared with other recommendation algorithms, the random forest recommendation algorithm has the advantages that the superiority of the random forest classification algorithm is fully understood, and the random forest classification algorithm is autonomously realized, so that the understanding of a recommendation system is deepened; and finally, key points causing algorithm deficiencies are researched from algorithm construction, and the algorithm is improved, thereby enabling the algorithm to have better efficiency.

An object-focused workflow system for processing a received object in accordance with a declarative workflow specification. The specification includes modules and attributes, where module execution results in the evaluation of attributes, and may include the initiation of a side-effect action performed by an external component. Whether modules are to be executed for a particular received object is determined by associated enabling conditions. Attributes may be evaluated in accordance with computation rules and a combining policy, where the computation rules specify how values are to be contributed to an attribute, and the combining policy indicates how those contributed values are combined in order to assign a value to the attribute. Tasks in the workflow system may be executed eagerly in order to improve the performance of the workflow system. The eager evaluation of tasks includes the determination of whether such tasks are eligible for eager evaluation, and whether the tasks are unneeded or necessary for the processing of the received event. Workflows which satisfy described design properties allow for improved algorithms for the determination of the whether tasks are eligible, eager, and / or necessary. A graphical user interface is provided for displaying a representation of the evaluation status of the modules and attributes during workflow execution.

The invention discloses a deep belief network image recognition method based on Bayesian regularization and belongs to the field of artificial intelligence and machine learning. The deep belief network plays a more and more important role in the field of digital detection and image recognition. The invention provides a deep belief network based on Bayesian regularization on the basis of the network sparsity characteristic and changes of connection weights to solve the problem of overfitting in the training process of the deep belief network. By applying Bayesian regularization to the network training process, balance between error decreasing and weight increasing is effectively adjusted. The classification experiment of a digital script database proves effectiveness of the improved algorithm. An experimental result shows that in the deep belief network, the deep belief network image recognition method can effectively overcome the overfitting phenomenon and improve accuracy of digital recognition.

The invention provides a network security verifying device based on an improved K-means algorithm. A data mining method is used for the extraction of network attack characteristics. Firstly, traffics entering and exiting a system are monitored through creating a height-controllable target range verifying environment, then abnormal traffics and abnormal behaviors which enter and exit the system are captured by layers by using a layering mechanism, and finally, collected information data are subjected to characteristic analysis by using an improved K-means algorithm, and therefore, the attack behaviors can be efficiently detected and a potential unknown attack in a network is accurately found.

The invention discloses a cascade reservoir optimal operation method based on the adaptive particle swarm optimization algorithm. According to the method, aiming at the defect of the particle swarm method in cascade reservoir optimal operation, fixed initialization improvement is conducted firstly on particle random initialization to enable the algorithm to have the possibility of approaching the optimal value at the beginning, large-scale dead zones do not exist, convergence speed is increased, and the stability of the algorithm is improved; then according to the group cooperation idea and the cluster ecological niche idea, an initialized group is dynamically divided into three subgroups, optimization and parameter selection are conducted on each subgroup in an adaptive mode according to the difference of particles, and in this way, the particle diversity is improved, the information exchange model is changed, and local optimum of the algorithm is avoided. According to the improved algorithm, the function problems of nonlinearity and multiple local minima can be well solved, and an effective and feasible solution is provided for cascade reservoir optimal operation.

The invention relates to an object detection method based on a cascade convolution neural network. The object detection method based on the cascade convolution neural network comprises steps of usinga convolution neural network to extract image characteristics, using a regional candidate network to generate a certain quantity of object candidate frames, using an optimized network to optimize a candidate framework, inputting the optimized object candidate frames into a detection network containing multiple classifiers to produce a preliminary detection result, using a binary classifier to perform re-detection on each kind of objects, and removing an error target to obtain a final accurate detection result. The object detection method based on the cascade convolution neural network uses anadvantage that the deep convolution network has a strong expression ability for a target, constructs a cascade convolution neural network for target detection, brings forward a new method of optimizing the target candidate frame and a strategy of eliminating an error detection sample, improves algorithm detection accuracy and can obtain a better target detection result.

The invention provides a KNN-based improved missing data filling algorithm, which comprises the steps of (1) improving a traditional multiple correlation coefficient inverse weighting method and calculating the importance of each attribute on a missing value-containing attribute by using an improved algorithm, deleting a few of attributes with relatively small correlation with a key attribute and carrying out streamlined operation on an attribute set to obtain a data sample set which only contains the streamlined attribute set; (2) comprehensively considering the advantages of the correlation between the attributes and the variability by using a mahalanobis distance, effectively predicting an uncertain factor-containing sample by combining a grey correlation analysis method and calculating K adjacent samples of a missing sample; and (3) giving entropy weight values to the attributes corresponding to the K samples according to the calculated K distance values and an entropy weight method and then calculating a final filling value by combining attribute values. According to the KNN-based improved missing data filling algorithm, the calculating complexity of the missing data algorithm can be reduced, the accuracy of the adjacent sample values is improved and the estimation accuracy of the data filing value is improved.

The invention discloses a smart phone-based indoor positioning and pedestrian tracking method. The method comprises the following steps: 1, preliminarily determining the position and the locus of a sensor in a phone; 2, fusing and preprocessing indoor information; 3, assisting the preliminarily positioned position information by using the fused information; and 4, calculating an accurate linear locus according to the fused information by using an improved equalization algorithm, and correcting the angle error of a pedestrian at a corner by using an improved algorithm in order to integrally realize accurate positioning and pedestrian locus tracking. The method solves the partial fusion problem of existing indoor position, integrates the functions of the smart phone, ingeniously fuses all indoor position through the improved algorithm, improves the positioning and locus accuracy of the indoor pedestrian, utilizes the functions of the daily necessity phone in life, and has the advantagesof low consumption and convenience.

The invention discloses an underground personnel positioning system based on a radio frequency identification technology. The system comprises tag card readers, electronic tags, a switch, an upper personal computer (PC), a central display screen and optical fibers; at least three tag card readers are arranged in underground laneways and are connected with the electronic tags which are worn by underground personnel through radio frequency electromagnetic waves; the conventional ranging algorithm is improved by adopting an electromagnetic wave propagation model and an attenuation index which are in accordance with an underground environment; distance between the underground personnel and each tag card reader is measured accurately by an improved ranging algorithm by the tag card readers; measurement results are transmitted to the ground upper PC after being collected by the switch; coordinate positions of the underground personnel are determined accurately by a trilateral positioning method according to the distance value of every three same identifiers by the upper PC; a positioning result is displayed on the central display screen in real time; and position reference is provided for ground management personnel and personnel who is about to go down for operating. By the improved system and the improved algorithm, the problem of coal mine underground positioning difficulties is solved effectively.

The invention relates to a real-time blood pressure monitoring and timely remote service device with an acceleration transducer and a system thereof. The real-time blood pressure monitoring can be used for dynamically monitoring physiological parameters of human bodies such as blood pressure, heart rate and the like in real time, a pressure-frequency conversion method is used for converting a pressure signal into a frequency signal which has strong anti-interference capacity, thus enhancing the anti-interference capacity of the real-time blood pressure monitoring. An oscillography-based blood pressure measurement improvement algorithm effectively optimizes use of an internal memory space. The acceleration transducer can judge motion state of a tested person and control startup of blood pressure measurement to enhance measurement efficiency and save power. A GSM module sends monitoring data to a remote service platform, the heart rate data of the tested person are analyzed, and the system compares and selects the blood pressure data according to a preset heart rate variation threshold, removes abnormal blood pressure data, and stores the useful data. An intelligent signal analysis and processing program automatically analyzes the blood pressure data and sends alarm information to the tested person at proper time.