39results about How to "Effectively find out" patented technology

The invention belongs to the technical field of information retrieval, and particularly relates to a music retrieval system based on audio fingerprint features. The system is composed of a preprocessing module, a feature extraction module, a reverse index module and a fine matching module. The preprocessing module mainly carries out audio signal conversion, resampling and filtering; the feature extraction module is used for representing audio files, wherein the audio fingerprint features are adopted to select the most stable point from a frequency spectrum as the feature point through twice screening based on dynamic threshold values, and each feature is represented by a dot pair; according to the reverse index module, the features are used as key words, reverse indexes are built according to the features of a song library, and the index result is returned according to the number of the same key words; according to the fine matching module, the sequential relationship of the audio features is combined, an improved editing distance is adopted as the similarity of two feature sequences, and therefore the index result is optimized. The music retrieval system based on the audio fingerprint features is suitable for the retrieval of a large number of songs, and can particularly conduct effective retrieval on record inquiry segments.

The invention relates to a certificateless key agreement method on the basis of a mobile Ad Hoc network, which comprises the following steps of: 1, initializing a system network; 2, constructing a network main private key, i.e. selecting a (t-1)th-degree polynomial to exchange necessary information at each node, enabling random t nodes or more than t nodes to mutually cooperate to obtain the network main private key and simultaneously obtaining a network public key; 3, generating key pairs for the nodes by utilizing the main private key in the network; 4, carrying out secrete communication among the nodes by utilizing the generated keys so as to implement the encrypting and decrypting process. The certificateless key agreement method has no problems of certificate verification and key escrow when implementing key agreement; meanwhile, dangerous nodes can be detected; moreover, a recommended encrypting scheme can have fewer encrypting items on the premise of ensuring security and the calculated amount and the communication traffic are reduced, and thus, the certificateless key agreement method has good practicality and security.

The embodiment of the invention provides a determination method and system of an equipment threshold value. The method comprises the following steps: obtaining a data sample of the equipment, wherein the equipment is from the same totality; judging whether the data sample complies with normal distribution or not, and if the data sample complies with the normal distribution, determining an expectation and standard difference of the data sample; and according to the expectation and standard difference of the data sample, calculating a significant difference threshold value of the data sample, and taking the significant difference threshold value as the threshold value of the equipment. The threshold value, which is subjected to substantial change, of the equipment is effectively found out, the fault hidden dangers of various pieces of equipment can be removed in time, and the precision and the adaptability of the equipment threshold value can be improved.

The invention relates to a waterproof electronic device with a test structure, a water leakage tester and a water leakage testing method. The waterproof electronic device with the test structure mainly comprises a test hole, a sealing structure and a fixing structure, wherein the test hole is formed on a shell of the waterproof electronic device, and is used for introducing a test fluid to perform a water leakage test; the sealing structure is arranged on the shell or the water leakage tester, and can be used for accommodating a sealing element to form seal between the waterproof electronic device and the water leakage tester in the water leakage test; and in the water leakage test, the fixing structure can be used for fixing the water leakage tester to the waterproof electronic device. The waterproof electronic device with the test structure, the water leakage tester and the water leakage testing method are applicable to the water leakage test at various stages such as a development stage, a production process, after-sale stage and the like, can effectively, rapidly and economically find water leakage positions to clarify reasons for water leakage and timely take remedial measures, and are practical.

An embodiment of the invention provides an abnormal medical cost judging method and an abnormal medical cost judging device, and relates to the field of medical technology. The abnormal medical cost judging method includes: acquiring clinic data information of all hospitals within a preset area, processing the clinic data information of all the hospitals according to a preset mixed-model algorithmso as to determine medicine proportion outliers of the hospitals, determining a seeming abnormal charge data information collection according to the medicine proportion outliers, and finally judgingthe hospitals charging abnormal medical costs according to the seeming abnormal charge data information collection and a preset medicine proportion range. The abnormal medical cost judging method andthe abnormal medical cost judging device are capable of judging whether treatment cost of a hospital within the preset area is abnormal or not by processing the clinic data information of the hospital, and have the advantages of practicability and accuracy.

The invention relates to a bill frame rapid positioning method based on color line segment detection and a system thereof. The method comprises the steps of inputting a bill image, scaling the image in proportion, detecting color line segments, merging the line segments, screening the line segments, building a maximum quadrangle and outputting a vertex. The method has the advantages that the colorbills can be recognized, the method is not affected by gray scale conversion, the vertexes of the bills are accurately and rapidly positioned, and the effective bill areas are cut out.

The invention discloses a static calibration method for camera external parameters. The method comprises the following steps: S1, setting a plurality of marking points; S2, converting the actual GPS position of the mark point into an IMU coordinate system taking the IMU as an origin of coordinates to obtain a point P1; S3, converting P1 into a camera coordinate system taking a camera as an originof coordinates by rotating the translation transformation matrix to obtain P2; S4, the P2 is converted into an image coordinate system through an internal reference matrix of the camera device, and projection coordinates of the point on the image are obtained; s5, calculating a re-projection error according to the constructed re-projection error function; and S6, optimizing the rotation translation transformation matrix, and repeating the steps S3 to S5 until the re-projection error is lower than a specified threshold. According to the method, a loss function is constructed through GPS coordinates of mark points and pixel coordinates of the mark points in a camera device to finally obtain an Euclidean transformation relation between the GPS coordinates of the mark points and the pixel coordinates of the mark points, namely a rotation and translation transformation matrix. According to the method, the calibration of the external parameters of the camera can be completed in a static state without reconstructing collection equipment and a vehicle.

The invention belongs to the technical field of routing node positioning, and particularly relates to a routing node importance sorting method and system based on routing characteristics, which are used for searching backbone nodes in a target network area, and comprise the following steps: acquiring path information among nodes in the target network area through network measurement, and constructing a network topology; pruning the network topology based on the stability of the path between the network topology nodes to obtain a topology network structure only containing a stable path; counting the number of stable paths between adjacent nodes in the topological network structure, and endowing each corresponding edge in the topological structure with weight according to a counting result; and sorting the target network routing nodes according to the weighting degree centrality. According to the method, the processed data volume can be effectively reduced, the effect of the edge in the network in actual communication can be remarkably distinguished, and accurate positioning is facilitated so as to discover more backbone nodes.

The invention relates to a reproducing apparatus, a reproducing method and a program. Via easy change of position (place) of partial reproducing of content information, the content required by a user can be effectively found, wherein, the reproducing apparatus 100 includes: a storage unit 120; a display control unit 115; a designating control unit 111; a position determining unit 112; an object determining unit 113; and a reproduction control unit 114. The storage unit stores information of each of a plurality of contents. The display control unit controls to display, on a display unit 130, content identification images corresponding to the plurality of contents respectively. The designating control unit 111 designates any one of the content identification images. The position determining unit 112 determines a reproducing position for the content based on a designating position for a display range of the content identification images which is designated when any one of the content identification image is designated. The object determining unit 113 determines the content as the reproducing object; and the reproduction control unit 114 reproduces the determined content from the determined reproducing position.

The invention discloses a sparse representation method for the on-line data collection of power, and the method comprises the following steps: (1) building a sparse coding mode, and initializing a zero vector; (2) randomly sampling a data sample for iteration; (3) updating a sparse coding coefficient; (4) calculating an approximation error; (5) judging the algorithm convergence. The method is used for the research of big data of power based on a random gradient descent algorithm, can quickly and effectively process data, greatly improves the solving efficiency, also can effectively solve a sparse coding coefficient, achieves the compression of big-data sparse coding, reduces the storage space, and reduces the storage requirements for hardware.

The invention discloses a composite insulator defect detection method. The method comprises the following steps: T1, acquiring an infrared image of a composite insulator string by using an infrared instrument; T2, reading the maximum temperature value of each composite insulator in the composite insulator string through the infrared image, wherein i belongs to (1, n), and n is the number of the composite insulators in the composite insulator string; T3, solving a corresponding curved surface score Ki according to the maximum temperature value of each composite insulator, wherein S is the area of the side surface of the single composite insulator umbrella skirt sheath; T4, obtaining the difference value [delta]Ti of the corresponding curved surface fractions of every two adjacent composite insulators, wherein [delta]Ti=Ki+1-Ki; and T5, judging whether defects exist in the composite insulator string or not according to the absolute values of all [delta]Ti and / or the positive and negative values of all [delta]Ti. According to the method, the curved surface integral technology is utilized, the temperature detected by infrared is integrated along the side surface of the composite insulator, and the integral result of the degraded composite insulator is greatly increased compared with the integral result of a normal insulator.

The present invention provides a method for extracting entity address information in a text context, which collects a collection of web pages containing entity names, calculates the initial correlation degree and corrects the initial correlation degree to obtain the final correlation degree, and finally ranks according to the correlation degree , and return the top results to the user; the present invention can effectively find out the address information related to the physical space specified by the user from the addresses included in the web page, and assist the user to locate.

The invention discloses an intelligent residue removing device for a precision casting product, and relates to the technical field of casting product treatment. The intelligent residue removing device comprises a pouring device used for cleaning residues, wherein a torsion device used for clamping the product and driving the product to rotate is arranged at the upper end of the pouring device, a surface grinding device and a hole grinding device which are used for moving for grinding are arranged at the upper end of the torsion device, the surface grinding device and the hole grinding device have the hard grinding function and the soft grinding function, the two grinding surfaces can be switched, and a telescopic device used for grinding holes in the product is arranged on the hole grinding device. According to the device, the product needing to be ground is ground in all directions, two grinding heads are arranged to deal with different residues on various products, and cleaning is more comprehensive.

The invention discloses a deep learning model poisoning attack detection method based on mutual information. The method comprises the following steps: (1) obtaining a sample set and a to-be-detected deep learning model; (2) pre-training a deep learning model; (3) generating a poisoning model pool and a trigger sample pair; and (4) judging whether the deep learning model is poisoned or not by utilizing mutual information. The invention also discloses a deep learning model poisoning attack detection device based on mutual information, which is used for implementing the method. The method has good applicability, can effectively judge whether the model is poisoned or not and find out a poisoned target class, and can obtain a better detection effect.

The invention discloses a dynamic calibration method for camera external parameters. The method comprises the steps that S1, in the slow running process of a vehicle, one or more marking points are set, and the position of a corresponding GPS is obtained; S2, projecting the one or more mark points into an image acquired by the vehicle camera device through a transformation matrix M [T | t] to obtain theoretical projection coordinates of the one or more mark points on the image; and S3, constructing a loss function related to T and t, and calculating an error. A loss function related to a rotation and translation transformation matrix T and time delay t is constructed between coordinate observation values of marking points in a camera device and coordinate estimation values obtained througha series of transformation, and calibration of T and t can be completed at a time.

In order to carry out low-maintenance-cost and convenient-to-implement tests on radar reflection conditions of an aircraft cockpit command and control system under interference conditions, the invention discloses an aircraft cockpit command and control system simulation test device and an external device thereof. The simulation test device comprises an aircraft radar (1), a shielding cage (2), a test board (3) and a plurality of test wires (4). The external equipment comprises a first electromagnetic interference generation unit (5) and a second electromagnetic interference generation unit (6), the test board (3) is arranged in the shielding cage (2), the aircraft radar (1), the first electromagnetic interference generation unit (5) and the second electromagnetic interference generation unit (6) are arranged outside the shielding cage (2), each of the plurality of test wires (4) comprises two wiring terminals, one end of each wiring terminal is connected with the test board (3), and the other end of each wiring terminal is connected with a flight attitude control signal output end of an aircraft cockpit command and control system.

The invention relates to a receiving unit for searching for at least one unused transmission channel in a communications device in a network of distributed stations and / or for monitoring transmission channels such that specific messages can be determined for that communications device, and to a method for its use, in particular in a local broadband network or a digital broadband home network. The receiving unit comprises a filter unit (5), which has a broadband reception characteristic in a monitoring mode, such that signals from a number of transmission channels can be received in parallel; a digital filter device (7) for subdividing a broad filter band, which has passed through the filter unit (5), into a number of component signals which each have a narrower bandwidth than broadband signal and can each be associated with one of the number of transmission channels; and an evaluation device (9) in order to determine at least one signal strength for each of the number of component signals so that it is possible to decide which of the number of transmission channels is or are being used and / or which of the number of transmission channels is or are not being used.

The invention relates to a method and device for recognizing associated texts, and belongs to text processing in the natural language processing technology.The method comprises the steps of determining a first target text and a second target text; obtaining the text similarity of the first target text and the second target text; identifying a character type to which each character belongs in the first target text, and generating a first structured text according to an identification result; identifying a character type to which each character belongs in the second target text, and generating a second structured text according to an identification result; based on the first structured text and the second structured text, determining the text structure similarity of the first target text and the second target text; and determining an association relationship between the first target text and the second target text according to the character text similarity and the text structure similarity. The mutually associated target texts can be effectively found out.

The invention relates to the field of integrated circuit design and discloses a parallel integration method and system for integrated circuit design. The parallel integration method comprises the steps of: simultaneously generating a plurality of functional equivalent subgraphs of specific-mode subgraphs in an original CDFG (Control Data Flow Graph) by using the parallel processing capability in a computer technology, combining the functional equivalent subgraphs of each specific-mode subgraph into a plurality of CDFGs, carrying out parallel treatment and optimization on the plurality of CDFGs, simultaneously generating a plurality of low-level hardware circuits, and finally, selecting and determining the low-level hardware circuit with the optimal performance for one time. And therefore, the integration time required by the plurality of CDFGs is only one cycle time of the traditional integration flow, the time required for integration is greatly shortened, and in addition, as the combination of all the functional equivalent subgraphs is structured into a special CDFG to be optimized independently, the optimal circuit finally determined through parallel integration via parallel search is the optimal result which can be found out in all solution spaces.

The invention discloses a detection method of a carbon-encapsulated material for a carbon source. The detection method is characterized in that detection is accomplished through the following steps. The detection method has the advantages that carbon source materials of low densities in the carbon-encapsulated material and sample materials are effectively separated, meanwhile carbon sources not reacted in the material are simply and effectively found out by virtue of a scanning electron microscope, and an effective basis is provided for the performance judgment of the carbon-encapsulated material.

A pain-monitoring device and method according to an embodiment of the present disclosure is a pain-monitoring device for obtaining information related to a pain of a user and providing a service of monitoring the pain of the user, includes: an input unit sensing pain-related input of the user; a display unit outputting pain-related information; and a control unit controlling the input unit and the display unit, in which the input unit includes: a touch input unit sensing touch input of the user; and a pressure measurement unit measuring pressure input of the user, and the control unit creates user-input pain information in accordance with the sensed pressure input and touch input.

The invention discloses a scratch detection device. The scratch detection device comprises a box body, an image acquisition unit, an illumination unit and a control unit, wherein front, rear, left andright side plates, a top plate and a bottom plate are arranged on the box body; a first air tunnel and a second air tunnel are fixedly arranged at middle positions of the left and right side plates respectively; a third air tunnel is arranged at the middle position of the bottom plate; the third air tunnel is vertical to the bottom plate; the focal point of extending lines of the first air tunnel, the second air tunnel and the third air tunnel is a point A of a material to be detected; the speed of enabling the material to be detected to reach the points A through the third air tunnel is 0. The device can be used for effectively finding out materials with scratches, the production efficiency is improved and the labor cost is reduced.

The invention discloses a big data speech classification method. The big data speech classification method comprises the following steps of 1 collecting speed samples as a training set, 2 looking for a big data speech classification optimized frequency spectrum matrix, 3 conducting spectral analysis on unmarked data, and 4 adopting frequency bands to classify big data speeches according to frequency spectrum data. By means of the big data speech classification method, feature information of different speech data can be effectively found under the big data situation, accordingly various relevant data are effectively classified, storage cost in the training process is effectively reduced, and the classification accuracy is higher than the accuracy in the prior art.

The invention discloses an implementation method and device for a code hosting warehouse based on a block chain and a medium, and aims to solve the technical problems that an existing code warehouse is easy to fail to lose data and is easy to insert malicious codes during deployment. The method comprises the steps of obtaining preset block chain information uploaded by a first user, and creating a block chain based on the preset block chain information; deploying a code developer as a second user on the block chain node; receiving a smart contract code input by a first user and writing the smart contract code into the block chain; uploading the smart contract code to a second user through the block chain node; when a modification code input by a second user is received, brining the modification code automatically into the smart contract code, and verifying whether the modification code accords with the agreement of the smart contract code; and when the agreement is met, writing the modification code into the block chain. By means of the method, distributed storage of the codes is achieved, and the problems of code static errors, malicious code writing and the like are avoided.

In order to test the radar reflection of the aircraft cockpit control system under interference conditions with low maintenance cost and easy implementation, the invention discloses an aircraft cockpit control system simulation test device and its external equipment. The simulation test device includes: aircraft radar (1), shielding cage (2), test board (3), many test wires (4), described external equipment comprises the first electromagnetic interference generation unit (5) and the second electromagnetic interference generation unit (6), so The test board (3) is arranged in the shielding cage (2), and the aircraft radar (1), the first electromagnetic interference generation unit (5) and the second electromagnetic interference generation unit (6) are arranged in the shielding cage (2 ) outside, the plurality of test wiring (4) all includes two terminal blocks, one end of each terminal block is connected with the test board (3), and the other end is connected with the flight attitude control signal output end of the aircraft cockpit command and control system.

The invention relates to a bill frame rapid positioning method based on color line segment detection and a system thereof, comprising: inputting a bill image, scaling the image in proportion, detecting color line segments, merging line segments, screening line segments, building the largest quadrilateral, and outputting vertices. The beneficial effect of the invention is that the colored bills can be identified without being affected by the gray scale conversion, the vertices of the bills can be accurately and quickly positioned, and effective bill areas can be cut out.

The invention discloses a user entity behavior analysis method based on density peak value adaptive clustering, which comprises the following steps of: 1, extracting access user data of an important data server within a certain time period, and constructing a data graph according to access times and access behaviors; 2, calculating the local density of the user data points in the step 1 and the distance between the user data points; 3, taking a point before the inflection point as a cluster center point; 4, analyzing user behavior types in each data point cluster; 5, analyzing abnormal behavior information of the user according to the user behavior baseline constructed in the step 4; and step 6, screening a suspicious personnel list according to the information. According to the user entity behavior analysis method based on density peak value adaptive clustering, the user entity behaviors can be effectively analyzed through setting of the steps 1-6, and suspicious users can be found in time.

The invention relates to an e-commerce platform potential customer recommendation method and system based on comprehensive similarity. The method comprises the following steps: selecting a motif, finding out all instances of the motif, constructing a high-order adjacency matrix of customers, then obtaining a customer sequence, obtaining representation vectors of the customers, and further calculating the contact closeness between the customers; calculating the similarity between the features carried by the clients; weighting the connection closeness and the characteristic similarity to obtain the comprehensive similarity of the clients, and further obtaining a KNN graph; selecting the customers with the highest core from the customers which are not distributed with the groups as core customers; taking the selected core customer as an initial group, and sequentially adding the customer which enables the group fitness function increment to be maximum in group neighbor customers into the group; and core customer selection and group expansion are repeated until all customers in the customer network of the e-commerce platform belong to the groups. According to the invention, the group existing in the e-commerce platform client network can be effectively mined, and the commodity which is purchased most frequently is recommended for the group.