1418 results about "Evaluation function" patented technology

A cost estimation tool, preferably a handheld computer, prices and evaluates the financial impact of new residents' prior to admission. Immediate access to this information and built-in therapeutic substitution and clinical advisories provides the opportunity to optimize pharmaceutical regimens prior to admission. In addition, an abbreviated Resource Utilization Group (RUG) evaluation function provides a rapid assessment tool to project reimbursement under the Medicare Prospective Payment System (PPS). A user may enter a complete drug regimen and estimate the costs for the regimen. Similarly, managed care reimbursement from managed care organizations (MCOs) to be analyzed. Also, drug cost may be estimated through a MEDICAID Preferred Drug List database that identifies potential Non-Preferred drugs that may be non-compensable. The present invention further transmits results wirelessly to health care providers.

Embodiments of the invention describe thermostats that use model predictive controls and related methods. A method of controlling a thermostat using a model predictive control may involve determining a parameterized model. The parameterized model may be used to predicted ambient temperature values for an enclosure. A set of radiant heating system control strategies may be selected for evaluation to determine an optimal control strategy from the set of control strategies. To determine the optimal control strategy, a predictive algorithm may be executed, in which each control strategy is applied to the parameterized model to predict an ambient temperature trajectory and each ambient temperature trajectory is processed in view of a predetermined assessment function. Processing the ambient temperature trajectory in this manner may involve minimizing a cost value associated with the ambient temperature trajectory. The radiant heating system may subsequently be controlled according to the selected optimal control strategy.

For an information terminal to be operated by users for collecting predetermined pieces of information from remote information devices by free-space optical communication, the present invention provides a technique for suppressing the power consumption of the information terminal by minimizing the amount of calculation performed to collect the aforementioned information. According to the present invention, each information device emits ID light on which a low-frequency pilot signal is superimposed. The information terminal captures a series of frames of images including the ID light and locates the ID light within the images by the following steps: (1) creating multiple levels of binned images having different resolutions for each frame of the image; (2) calculating an evaluation index for each pixel within a target range of the binned images at each level, from the lowest to the highest resolution, where the target range is narrowed every time the process switches over to a lower level. In (2), the evaluation index is calculated by an evaluation function including fast Fourier transformation performed throughout the series of frames of images. The evaluation index thus calculated is compared with a threshold to determine whether the pixel concerned is receiving ID light. The present technique significantly reduces the number of pixel to be analyzed and evaluated, thereby decreasing the total number of arithmetic operations to be performed using the evaluation function. Thus, the power consumption is suppressed.

The present invention relates to an integrated video camera device and self-adapting automatic focus method. The method includes: using climbing method to a curve composed of an evaluation function evaluating the definition and the definition evaluation values of multiple images acquired continuously, selecting a focus method base on the zooming situation, searching the peak of the curve and driving the lens to reach the focusing position corresponding to the peak of the curve, performing real time motion detection to the lens scenes upon searching the peak, driving the lens motor to startup a new cycle of automatic focus process when a certain degree of fuzzy is judged to appear on the lens scene or the moving scene varies and at a non-focus status. The invention has beneficial effects that the device and the method can effectively determine the focus direction in a variety of zoom magnifications, can 'recognize' images at fuzzy status and accelerate them out of the fuzzy status, and overcomes the problems of the traditional focus method that it is difficult to judge focus direction in the fuzzy region under high magnifications, and the traditional focus method has slow focus velocity and repeated oscillation.

In order to acquire a suitable fingerprint image by correcting an elongated fingerprint image, a line sensor acquires fingerprint image as a plurality of line-shaped images. A computation unit computes a similarity value by use of an evaluation function for evaluating the degree of similarity between the line-shaped images. The similarity value represents the degree of similarity between a first line-shaped image and a second line-shaped image which serve as a similarity evaluation target and are included in the plurality of the line-shaped image. A compression unit compresses the first line-shaped image and the second line-shaped image when the similarity value is equal to or larger than a predetermined threshold value to generate a new line-shaped image. A generation unit generates the entire fingerprint image by combining the new line-shaped image with the other line-shaped images.

A fully computerized musical performance teaching system comprises a computer apparatus having a keyboard and a display and is connectable to an electronic musical instrument. An application program provides the computer with various functions such as of judging student's skills, selecting music pieces for practice, training the student in performance and evaluating the student's progress through the teaching course. The judging function is to judge the student's performance skill from the student's responses to the subjects which the computer presents. The selecting function is to select a piece of music for practice according to the judgment results and the student's wishes and also to select a tutoring manner from among several prepared manners. The training function is to plan a practice schedule to meet the student's performance skill and wishes and to let the student practice with proper music pieces. The evaluation function is to evaluate and mark the student's performance, and reflect the evaluation results and marked points in the progressing rate and the presented subjects and also in adjusting the judgment results.

For an information terminal to be operated by users for collecting predetermined pieces of information from remote information devices by free-space optical communication, the present invention provides a technique for suppressing the power consumption of the information terminal by minimizing the amount of calculation performed to collect the aforementioned information. According to the present invention, each information device emits ID light on which a low-frequency pilot signal is superimposed. The information terminal captures a series of frames of images including the ID light and locates the ID light within the images by the following steps: (1) creating multiple levels of binned images having different resolutions for each frame of the image; (2) calculating an evaluation index for each pixel within a target range of the binned images at each level, from the lowest to the highest resolution, where the target range is narrowed every time the process switches over to a lower level. In (2), the evaluation index is calculated by an evaluation function including fast Fourier transformation performed throughout the series of frames of images. The evaluation index thus calculated is compared with a threshold to determine whether the pixel concerned is receiving ID light. The present technique significantly reduces the number of pixel to be analyzed and evaluated, thereby decreasing the total number of arithmetic operations to be performed using the evaluation function. Thus, the power consumption is suppressed.

The invention discloses a knowledge graph representation learning method. The method comprises the following steps: defining the correlation between entity vectors and relation vectors in a relation triple (head, relation, tail) by utilizing a translation-based model between the entity vector and the relation vector; defining the correlation between entity vectors and feature vectors in a feature triple (entity, attribute, value) by utilizing a neural network classification model; and correlating the entity vectors, the relation vectors and the feature vectors with one another through an evaluation function, and minimizing the evaluation function to learn the entity vectors, the relation vectors and the feature vectors so as to achieve the optimization aim. By adopting the method disclosed in the invention, the relation among the entity, the relation and the feature can be accurately represented.

When a plurality of faces are arranged in order of distance, a three-dimensional effect evaluation value r for each face is determined on the basis of a three-dimensional effect evaluation function Fl having a point A which internally divides the distance from the front side to the rear side in the depth direction at 1:2 as the vertex. A face with the maximum three-dimensional effect evaluation value r, serving as a face with the highest priority, is determined to be a main object. Then, three-dimensional processing is performed on a plurality of images such that the parallax of the determined main object is the minimum, thereby generating an image for three-dimensional display.

The invention discloses an indoor AGV (Automated Guided Vehicles) path planning method based on an improved A* algorithm. The method comprises the following steps: firstly, constructing an environmental map by adopting a grid method, and expanding the grid map according to the size of an AGV; then, optimizing a search strategy of a traditional A* algorithm, introducing a multi-value map model, cost function evaluation and path inflexion smoothing, and adding safety factors as path evaluation indexes; and finally, adding a starting point and a target point of the AGV to the grid map, adopting an improved distance evaluation function, taking the node with the smallest cost function as a node of the next step, gradually planning a running route of the AGV, and finally displaying the running route in an electronic map. Simulated analysis verifies that the AGV path planning based on the improved A* algorithm is higher in search route safety factor, smoother in path and higher in search speed.

The invention provides an intelligent car hailing system and a car hailing method based on a block chain technique. The system provides demand discovery, transaction matching, settlement payment, and credit evaluation functions for multiple participants. A platform discovers a traveling car hailing demand of a passenger and a vacant passenger carrying demand of a driver, and the demands are stored to the block chain; a reasonable matching algorithm is adopted to match the passenger and the driver, and the transaction information is stored to the block chain; after the transaction, transaction settlement payment is carried out, and the transaction result is stored to the block chain; and a mutual evaluation function is provided for the passenger and the driver, and the evaluation information is stored to the block chain. The decentralized car hailing platform of the block chain technique is used to solve the problems of price control, market monopolization, data unfairness and benefit imbalance in the current car hailing market can be solved, and a healthy car hailing market is built.

An appliance measurement interface inputs an appliance measured value from each appliance. The appliance measured value includes an electric energy consumption, a usage intensity and a usage frequency of the appliance. An appliance information count unit calculates an average of the appliance measured value in an evaluation span, and outputs an appliance information as the average of the appliance measured value. An evaluation unit acquires a first appliance information in a first evaluation span as an evaluation object period and a second appliance information in a second evaluation span as a comparison object period from the appliance information, and calculates an evaluation value by applying the first appliance information and the second appliance information to an evaluation function. A display unit displays an evaluation result of an energy saving behavior based on the evaluation value.

A system and a method are provided for evaluating a voice performance in order to recognize a long-distance voice. The system implements a voice performance evaluation function for long-distance voice input in a robot. Particularly, in robots including a network robot, it is required to normally perform voice recognition so that a speaking subject and a surrounding situation can be recognized by a robot. Accordingly, in order to obtain the most optimal voice quality, it is important to find a noise removal algorithm through an optimal hardware configuration and an optimal combination of the optimal hardware configuration and software. Therefore, a method for finding a noise removal algorithm appropriate for each of cases, including one case where a distance from a speaking subject is fixed and another case where a distance from a speaking subject changes. As a result, the most optimal voice quality can be obtained regardless of a noise environment even when the speaking subject is a long distance away from the robot.

The invention relates to the technical field of automatic focusing of microscopes, in particular to an image processing based fast automatic focusing method of a microscope. In the method, the automatic focusing process is divided into a coarse focusing stage and a fine focusing stage; different focusing evaluation functions and different extremum searching strategies are adopted on different stages, specifically, on the coarse focusing stage, the gray scale different evaluation function and a fast hill-climbing backtracking search algorithm are adopted, and decision-making factors are introduced to prevent the search algorithm from falling into local extrema; and on the fine focusing stage, the Laplace operator based evaluation function and the small-step searching strategy are adopted. The method provided by the invention better solves the problem that the traditional algorithm is slow in focusing speed, easily falls into local extrema and is not high in focusing accuracy, and has strong real-time property and high reliability and practicability.

Machine learning for learning a characteristic amount of each pixel in a plurality of sample images, each including a particular region with a known contour, is performed in advance to obtain an evaluation function capable of evaluating whether or not each pixel is a pixel representing the contour based on the characteristic amount thereof. An arbitrary point is set within a particular region in an input image, and a discrimination region including the particular region is set in the input image with reference to the arbitrary point. A characteristic amount is obtained from each pixel within the discrimination region, then based on the characteristic amount, an evaluation value indicating whether or not each pixel within the discrimination region is a pixel representing the contour is calculated using the evaluation function, and the contour of the particular region in the input image is determined based on the evaluation value.

The invention provides an automatic focusing method of a camera. The automatic focusing method is characterized in that respective treatment is provided for selection of focusing windows correspondingly to normal illumination and low illumination, and on the basis without increasing excessive calculation, the inclusion of information at the upper half part, the left side and the right side of a picture in the focusing process is considered. A focus-adjusting search control strategy adopts an improved hill-climbing algorithm and comprises the steps of: firstly carrying out rough adjustment, then finding an extreme point by function fitting, driving a motor to reach near the extreme point, and then adopting a small-step method to carry out fine adjustment near the extreme point by the hill-climbing algorithm so as to enable the focusing to be more accurate. For the evaluation of the image definition, a Brenner function is adopted as an evaluation function of automatic focusing and is suitable for quickly finding the clearest point of the image in a manner of matching with the focusing process.

The invention belongs to the technical field of vision measurement, in particular to an automatic focusing method. In the method, an idea, which comprises that an image definition evaluation function is designed on the basis of improved discrete cosine transformation (DCT) to represent detail information of the image and highlight high frequency energy, and the relative definition is brought in by consideration of a close relation between the definition of the image and the brightness and the gray scale of the image, is provided. Meanwhile, in the method, a scale-variable window selecting scheme is provided on the basis of analyzing the influence from the selection of a focusing window on a focusing result. Experimental results show that the effect of the improved image definition judgment is improved greatly; in the actual application of the method, a partial extreme value of the image definition judgment is hard to avoid, so, by combining return information of a lens, an improved climbing method search strategy of total-range rough adjustment large step length focusing and partial refine adjustment small step length focusing is provided; and in the method, a scheme of starting automatic focusing when a shooting scene is changed or the lens is zoomed is provided.

A stereo image processing apparatus is provided. The apparatus comprises an imaging device for taking a pair of images of an object for stereo image processing purpose; a stereo processor for correlating the pair of images for each small area thereof to thereby perform stereo matching processing for identifying corresponding areas, and obtaining an evaluation function as a result of said stereo matching processing, distance information to said object obtained based on said evaluation function, and reliability information indicating a reliability of said distance information; a parallax corrector for setting an adaptive search range around a small area in question according to said reliability information, obtaining a corrected evaluation function by correcting said evaluation function obtained for said small area in question based on said evaluation function obtained for a respective small area within said adaptive search range, and correcting said distance information based on said corrected evaluation function obtained.

The invention discloses a Chinese domain term recognition method based on mutual information and a conditional random field model. The Chinese domain term recognition method includes the following steps: (1) gathering domain text corpus and marking all the punctuations, spaces, numbers, ASSCII (American Standard Code for Information Interchange) characters and characters except Chinese characters in the corpus; (2) setting character strings and computing the mutual information values of the character strings, (3) computing the left comentropy and the right comentropy of every character string, (4) defining character string evaluation function, setting evaluation function threshold, computing the evaluation function values of every character string, determining that every character string is a word, comparing in sequence the evaluation function value of the former character with the evaluation function value of the latter character in the character string and segmenting character meaning character strings one by one, (5) utilizing conditional random fields to train a conditional random field model and recognizing domain terms with the conditional random field model. When the Chinese domain term recognition method is used to recognize terms, the data sparsity of legitimate terms is overcome, the amount of calculation of conditional random fields is reduced, and the accuracy of the Chinese domain term recognition is improved.

A control apparatus for a radiotherapy irradiation system in which a calculation time required for calculation to be performed after restrictions are modified is shortened. The control apparatus includes a unit for setting an initial solution indicating an initial irradiated beam weight, a unit for inputting restrictions according to a prescription, and a unit for setting an evaluation function under the restrictions. The control apparatus further includes a unit for iteratively calculating the evaluation function until the evaluation function is minimized and determining an irradiated beam weight optimizing irradiation, and unit for outputting an optimal solution indicating an optimal irradiated beam weight to the irradiation system. After restrictions are modified, a new evaluation function is set under the modified restrictions. An irradiated beam weight is then determined for optimizing irradiation. At this time, the initial solution setting unit adopts the previous optimal solution indicating an optimal irradiated beam weight, which is obtained before the restrictions are modified, as the initial solution indicating the initial irradiated beam weight.

The invention provides a path programming method based on an A* algorithm capable of searching continuous neighborhoods. The path programming method comprises the following steps: establishing an environment model by using a grid method according to the geometrical size of an existing barrier, and treating a circular barrier by considering UUV as a mass point, taking the maximum width of the barrier as a diameter and taking the center of gravity of the barrier as the original point; acquiring the grid size l according to the information of the barrier; determining an evaluation function f (x) of the A* algorithm according to an established grid map; determining evaluation cost h(y) of an optional point y according to features of neighborhoods and the A* algorithm; and finding out a node of the minimum evaluation function fmin of the neighborhoods according to the evaluation function of the A* algorithm capable of searching the continuous neighborhoods as a next way point, and implementing a UUV sea route plan step by step. By the path programming method based on the A* algorithm capable of searching continuous neighborhoods, the problem that an existing UUV path planning method has path smoothness difference and a non-shortest path in a global environment is solved.

A short-term planned route from a start position that is a ship position at a first time up to an end position that is the ship position at a second time is designed based on a planned route, an estimated encounter marine phenomenon information, operation performance information measured on a ship, and a hull motion model of the ship, the estimated encounter marine phenomenon information being measured on the ship based on actually encountered marine phenomenon information. The short-term planned route makes a first evaluation function optimal, the first evaluation function containing: an index indicating an influence of a fluctuation portion between a planned position that is a ship position planned at the second time on the planned route and the end position; a fuel consumption index when the ship sails along the short-term planned route; and a safety index when the ship sails along the short-term planned route.

The invention discloses an underwater vehicle path planning method based on ocean current historical statistic information, comprising the following steps of: determining a sailing region, rasterizing the sailing region, generating an ocean current field in the sailing region by an ocean current historical statistic database, taking an electronic chart as an environment field to simplify and combine obstacles, islands and phytal zones in the sailing region and generate a prohibited area, storing ocean current information and prohibited area information according to grids, creating a path evaluation function, searching for an optimal path by a particle swarm optimization algorithm, outputting the optimal path and ending the path planning process. In the method provided by the invention, the ocean current field closer to a true value is generated by the ocean current historical statistic database, under the condition of taking full account of ocean current influence, the path evaluation function is designed based on safety, economical efficiency and smoothness, the particle swarm optimization algorithm is used as a path searching algorithm to perform global path planning for the underwater vehicle so as to plan an underwater vehicle sailing path which is closer to a practical sailing path.

A passive optical network (PON) in accordance with the invention uses a superframe having a number of subframes arranged in a two-dimensional array, wherein for normal data transfer subframes are allocated to each optical network unit (ONU) column by column, left to right, and within each column subframes are allocated from top to the bottom. Initially, for ranging, at least two subframes are allocated to an ONU, adjacent to a diagonal of the superframe (which may go from the top left corner of the superframe to the bottom right corner, or alternatively from the top right corner to the bottom left corner). In some embodiments, instead of allocating subframes column by column, the subframes are evenly spaced apart from one another, and conflicts with a previous allocation are resolved by adjustment, and optionally an evaluation function may be used to find an optimal allocation.

Interference applied to neighbor cells during terminal handovers can be reduced in wireless communication systems containing base stations having different cell sizes. A serving cell makes handover judgment criteria easier based on pre-acquired cell size information on its own cell and cell size information of neighbor cells in order to facilitate terminal handovers from a base station with a large cell size to a base station with a small cell size. The serving cell finds evaluation function values based on the propagation state (e.g. reference signal received power) between the terminal and target cell, and the interference and load information exchanged between the base stations, and selects the terminal for handover by comparing the evaluation function value with the handover criteria value. The system selects a target cell having a good effect in lowering interference such as a cell with large reference signal received power acquired from a terminal, to serve as the handover destination.

A radio-wave arrival-direction apparatus calculates a correlation matrix of received signals by correlation calculation between antenna elements, and calculates a noise spatial eigenmatrix, of which each row or column is an eigenvector belonging to a noise eigen-space, by eigenvalue factorization of the correlation matrix. The apparatus also factorizes a matrix including a product of the noise spatial eigenmatrix and a conjugated and transposed matrix of it to an upper or lower triangular matrix, using cholesky factorization. The apparatus calculates an angle evaluation value in a predetermined angle range of an arrival-angle evaluation function using the derived upper or lower triangular matrix, and determines an arrival angle based on the calculation result. A calculation amount in a variable angle range can be thus reduced without causing accuracy degradation of arrival direction, in an algorism requiring all angle sweep for arrival angle estimation of MUSIC method or the like.

The invention discloses a method for designing the formulation combination of tobacco leaf, which comprises the following steps: the design objective of cigarette products is determined; the requirement for the main chemical composition of the objective is determined; according to the experience, formulation personnel preliminarily select one representative tobacco leaves respectively from base cigarette, auxiliary material cigarette and filler; according to the similarity of the base cigarette and the compatibility of the auxiliary material cigarette and the filler, the rest tobacco leaf materials are selected; the final tobacco leaf materials used for the formulation are determined by the formulation personnel according to the storage rate of tobacco leaf, absolute degree of association and smoke panel test; the formulation cost is taken as evaluation function of fitness, and the proportion of the chemical composition, the main materials, the auxiliary materials and the filler are taken as constraint condition, the genetic algorithm is searched. The invention overcomes the deficiency that the prior art can not ensure the aesthetic quality, product style and flue gas index of the formulation of tobacco leaf to be in accordance with design requirements, and provides a method which supplies multiple formulation combinations that are in accordance with requirements, reduces the cost and lightens the workload of the formulation personnel, thus being widely applied to the field of cigarette production.