150results about How to "Improve error correction efficiency" patented technology

The invention provides a voice recognition error correction method and device based on artificial intelligence, and a readable medium. The method comprises the following steps: determining the targetfield of voice recognition; when the voice recognition original text has the original fragment having the smoothness which is not larger than the preset smoothness threshold value in the target field,performing the ED fragment recalling on the original fragment needing to be corrected, and obtaining a candidate fragment set of the original fragment; scoring each candidate fragment by means of thepre-trained fragment scoring model; according to the scoring of each candidate fragment, obtaining a target fragment corresponding to each original fragment in the target field by means of decoding,and obtaining a correction text for voice recognition in the target field. According to the technical scheme, the defects in the prior art are overcome, and the voice recognition error correction scheme is provided, the error fragment in the voice recognition original text can be corrected, and the accuracy of voice recognition is effectively improved.

The invention discloses a word input method and a word input system. The method comprises the following steps of: detecting misinput fragments existing in coded strings input by a user; compiling the misinput fragments according to legal character sets acquired in advance to acquire error correction candidate strings; and converting the error correction candidate strings into input texts. By the method and the system, errors can be more effectively corrected in the word input process; and the method and the system have wider application ranges.

Techniques for correcting gradient non-linearity bias in mean diffusivity measurements by MRI systems are shown and include minimal number of spatial correction terms to achieve sufficient error control using three orthogonal diffusion weighted imaging (DWI) gradients. The correction is based on rotation of system gradient nonlinearity tensor into a DWI gradient frame where spatial bias of b-matrix is described by its Euclidian norm. The techniques obviate time consuming multi-direction acquisition and noise-sensitive mathematical diagonalization of a full diffusion tensor for medium of arbitrary anisotropy.

Provided is a data transmission/reception method for simultaneously transmitting MBMS data from multiple cells through a HARQ method by allocating a shared feedback channel to user equipments of each cell and, when there is an error in the transmitted MBMS data, reporting feedback on a cell basis. The method for transmitting identical MBMS data from multiple cells to user equipments includes: a) transmitting control information including information on a service group identifier for the MBMS data and on a cell-specific shared feedback channel shared by user equipments of a cell to the user equipments through a control channel; b) transmitting MBMS data to the user equipments; c) when an acknowledgement is received through the shared feedback channel from at least one user equipment, transmitting radio resource information of retransmission data to the user equipments through a downlink control channel; and d) transmitting the retransmission data to the user equipments.

The invention provides a pinyin error correction method and system. The pinyin error correction method comprises the following steps: building a word bank, organizing all pinyin strings and reverse strings which can be retrieved by a user respectively, and storing the pinyin strings and the reverse strings in positive and negative ternary search trees; analyzing and detecting a retrieval string input by the user, judging whether the retrieval string can be split into a plurality of syllable strings or not, and acquiring the error type of the retrieval string, wherein the error type is classified into a legal error and an illegal error; performing legal error correction on a legal error retrieval string, performing illegal error correction on an illegal error retrieval string, and acquiring an error correction result by adopting inquiry operation of the positive and negative ternary search trees; computing the similarity of all pinyin strings and the retrieval string input by the user in the error correction result, acquiring K pinyin strings with highest similarity, and outputting the K pinyin strings. The pinyin error correction method provided by the invention has the advantages of simple and efficient design, high error correction speed and high accuracy.

The invention relates to a multiparty quantum key distribution system error correction method, which comprises the following steps: S1) calculating system error rate; S2) finding an optimal segment length sequence according to the system error rate; S3) carrying out loop iteration execution of the following steps: S3.1) carrying out data rearrangement; S3.2) carrying out segmentation on sender original key and distribution key according to the optimal segment length, calculating odd-even checksum of the segmented key of a sender and a receiver respectively, comparing operation results of the sender and each receiver to obtain the segmented keys, the operation results of which are different; S3.4) through a Winnow error correction algorithm, finding the error bit position; and S3.4) according to the error bit position, carrying out inversion on the 0 or 1 value at the corresponding position in a shared key of each receiver until reaching the largest iteration times and stopping cycle; and S4) checking the error correction result, if each receiver corrects error successfully, ending the operation, and if at least one receiver fails error correction, discarding the corresponding key. The multiparty quantum key distribution system error correction method is high in reliability and error correction efficiency, and can be suitable for multiparty quantum key distribution system error correction.

The invention relates to an error correction method for a multi-party quantum secret sharing system. The error correction method comprises the following steps: S1: calculating a system error rate; S2: according to the system error rate, finding the optimal segment length sequence; S3: performing the following steps through loop iteration: S3.1: rearranging data; S3.2: according to the optimal segment length, segmenting an original key of a transmitter and a shared key; calculating parity check sums of segment keys at a transmitting end and a receiving end respectively, and comparing operational results of the transmitting end and the receiving end to obtain the segment keys with different operational results; S3.3: finding error bit positions by a Winnow error correction algorithm; S3.4: according to the error bit positions, negating 0 or 1 values at corresponding positions in the original key of the transmitter, and stopping loop iteration till the maximum number of iterations is reached; S4: checking error correction results; if error correction is successful, ending operation; if the error correction is not successful, discarding the corresponding keys. The invention provides the error correction method which is high in reliability and high in error correction efficiency and is applicable to the multi-party quantum secret sharing system.

The invention provides a long text error correction method and device based on artificial intelligence, and a computer readable medium. The method comprises the following steps that: when the originalfragment of a nondedicated noun is present in a long text, according to a preset PT (Phrase Table) in the field of the long text, carrying out PT segment recalling on the original segment which needsto be subjected to error correction to obtain the candidate segment set of the original segment, utilizing a pre-trained segment marking model to independently carry out marking on each candidate segment in the candidate segment set, wherein the candidate segment set comprises a plurality of candidate segments; and according to the score of each candidate segment, through a decoding way, obtaining a target segment corresponding to each original segment from the candidate segment set of each original segment, which needs to be subjected to the error correction, of the long text, and therefore,obtaining the error correction text of the long text. Through the technical scheme of the invention, deficiencies in the prior art are made up, a long text error correction scheme is provided, wrongsegments in the long text can be corrected, and the editing quality of the long text is effectively improved.

The invention discloses a coding method, a decoding method, an error correction method and device. The coding method is suitable for executing RS coding so as to perform error correction on data including one symbol error code transmission in one clock period. The code word length of information code data input in one clock period is k, the code word length of code data output by the RS coding isn, the code word length of corresponding check code data is 2t, and 2t is equal to n-k, the length of one symbol is t. The coding method comprises the following steps: carrying out combination XOR oninput information code data and a preset check matrix to obtain check code data; combining the information code data with the check code data to obtain RS code data, and outputting the RS code data; wherein the check matrix has 2t rows and n columns, the check matrix comprises a first interval and a second interval, the first interval comprises k/t continuous t-row and t-column unit matrixes, andthe second interval comprises a 2t-row and 2t-column unit matrix. By adopting the scheme, the coding efficiency can be improved.

The present invention discloses a method of brightness correction for the electronic display, which consists of the following steps: take pictures of the electronic display; get the characteristic values of all the light-emitting components in the pictures; calculate the correction values of each light-emitting component with the characteristic values; correct the brightness of the display with the correction values. The present invention reduces the time cost of measuring the actual brightness values of the light-emitting components, and improves the efficiency of correcting the brightness uniformity of the electronic display.

The invention relates to the field of quantum secret key error correction, and especially relates to a method for carrying out cascading error correction for an error code in quantum secret key distribution postprocessing. The method comprises the steps: carrying out multiple error correction of a secret key block; finding the position of a corrected error code in a one error correction process after the error correction process is finished; finding the position of the error code before the random rearrangement through historical record information, and generating a new block with an odd number of errors at this moment, and carrying out the secondary error search of the new positioned block with an error code; carrying out the continuous backtracking of error correction, stopping the correction when the odd and even verification of all blocks of a transmitting end and receiving end in the record in the whole process are consistent, and entering the next process of error correction at this moment. The method carries out backtracking search in an error correction process, and carries out error correction with errors. The method employs the historical information to fully dig the interaction information of two communication sides, achieves an effect, which can achieved in a conventional protocol through a plurality of processes, in one error correction process, and greatly improves the error correction efficiency.

The invention relates to an ECC controller based on Reed-Solomon codes, provided with a coding register, a decoding register, a control register, a status register, an error correction logic unit, a data-partitioning logic unit and a control module. The data-partitioning logic unit transmits coding/decoding data to an Reed-Solomon code coding/decoding module in a symbol mode for coding/decoding processing while a CPU inputs the data to the coding register or the decoding register so that a single data flow finishes a data transmission process and a corresponding ECC code or error code-generating process and the ECC controller carrying out the parallel processing of data transmission and ECC code or error code operation is realized.

A method for manufacturing an electronic circuit board which contains an electronic circuit on a main surface of a glass substrate according to the invention sequentially performs a step of electrically inspecting the main surface of the glass substrate on which the electronic circuit is formed, a step of specifying positions and defect types of defects on the main surface of the glass substrate, a step of calculating reference point coordinates on the main surface of the glass substrate and correcting the coordinates, a step of extracting respective defects from an image around the defects and specifying a defect to be corrected in the extracted defects by referring to a defect existing range registered in advance for each defect type, and a step of cutting the specified defect. By this method, a foreign material adhering to the glass substrate is not erroneously judged as short-circuit defect causing short-circuit on the electronic circuit, and only defect actually requiring correction can be removed.

A memory device includes: an in-memory error correction code generating circuit suitable for generating an in-memory error correction code based on a data received from a memory controller during a write operation; a memory core suitable for storing the received data and the in-memory error correction code during the write operation; an in-memory error correction circuit suitable for correcting an error of the data which is read from the memory core based on the in-memory error correction code which is read from the memory core during a read operation; and a data transmitter suitable for transferring the data whose error is corrected by the in-memory error correction circuit to the memory controller during the read operation, and transferring the data which is read from the memory core to the memory controller during a read retry operation.

The invention discloses a text error correction method and device, and relates to the technical field of artificial intelligence. The specific implementation scheme is that the method comprises the steps: obtaining a to-be-corrected text and a field to which the text belongs, wherein the text comprises at least one word; for each word in the text, querying a recall word list of the domain according to the word, and judging whether a to-be-corrected word matched with the word exists or not, wherein the recall word list comprises a corresponding relation between the to-be-corrected word and a domain keyword; when the to-be-corrected word matched with the word exists in the recall word list, determining a domain keyword corresponding to the matched to-be-corrected word as a candidate word corresponding to the word; and in combination with the candidate words corresponding to the words in the text, performing error correction processing on the text to obtain an error-corrected text. According to the method, the number of recalled candidate words can be greatly limited, and the text error correction efficiency is improved.

The invention belongs to the field of machining lines of shafts and in particular to a device for detecting and qualifying diameter dimensions of shafts. The device provided by the invention comprises a positioning mechanism, a measuring component and a limiting mechanism, wherein the positioning mechanism is used for being positioned with a measured shaft; the measuring component can move linearly; the measuring component moves horizontally along a guide mechanism; a reference part is arranged on the measuring component; the distance between the reference part and the shaft line of the measured shaft is matched with the standard value of the radius of the measured part of the measured shaft; and the limiting mechanism is used for limiting the nearest position of the measuring component. Because in the invention, the error correction is realized by depending on the physical measurement in a form similar to a go/no-go gauge, the device has the advantages of good stability and high error correction efficiency and can effectively avoid the occurrence of production accidents.

The invention provides an error correction method used for improving the data error correction efficiency under a high error rate environment. The method comprises the steps as follows: S1, a data sending end builds a communication connection with a data receiving end; and the structure of a frame of data in a data physical layer sent by the data sending end is formed into a header, a CRC check code and a data sting; and S2, a data receiving end receives a data frame M from the data sending end; the CRC check code in the data frame M is checked, if the condition that an error code occurs in the data frame M is detected, information is fed back to the data sending end to perform data retransmission, the data receiving end further receives a retransmission frame N from the data sending end;during an actual application process, if the CRC check code passes the check, the combination of data difference parts of the data frame M and the retransmission frame N is correct data, if the CRC check code cannot pass the check, the other retransmission frame N+1 is waited continuously; according to data error correction in such mode, the data error correction efficiency under the high error rate environment can be significantly improved.

The invention relates to an encoding error correction method of power line power frequency communication. The method comprises the following steps: receiving a R (x) code, calculating a syndrome S (x) according to the received R (x) code, calculating an error position polynomial sigma (x) according to the syndrome S (x), solving a root of the polynomial sigma (x) to acquire an error position number, calculating an error pattern E (x) according to the error position number, and calculating a C (x) code with errors corrected through the R (x) code and the error pattern E (x). By adoption of the encoding error correction method provided by the invention, random errors and burst errors can be corrected, the encoding error correction efficiency is effectively improved, and the accuracy of the power line power frequency communication is effectively improved.

The invention discloses a Jielin code error correction optimization method and device. The method comprises the steps: S100, obtaining to-be-coded binary data, carrying out the bit-by-bit calculationof the to-be-coded binary data through a first constraint condition based on a weighted probability model to obtain the coded data; then obtaining the bit length of the binary data to be coded and thefirst total number of symbols 0, and sending the coded data, the bit length and the first total number to a channel for data transmission; and S200, receiving the coded data, the bit length and the first total number from the channel, decoding the coded data based on the weighted probability model, checking a decoding result through the first constraint condition, and performing forward error correction processing on the decoding result. According to the invention, the data transmitted in the channel is linearly encoded and decoded through the weighted probability model, so that the requirement on hardware resources is low, the error correction speed is improved, and the error correction efficiency is further improved.

A memory system includes: a first error detection circuit suitable for generating a first error detection code using host data and a host address which are transferred from a host; a second error detection circuit suitable for generating a second error detection code using system data including one or more host data, a logical address corresponding to one or more host addresses, a physical address corresponding to the logical address and one or more first error detection codes; a third error detection code suitable for generating a third error detection code using the system data, the one or more first error detection codes and the second error detection code; and a first memory suitable for storing the system data, the one or more first error detection codes, the second error detection code and the third error detection code.

The invention discloses a cross polarization error correction method for a microwave radiometer. The defects of more correction steps, large calculated amount and heavy dependence on ground test parameters in an existing cross polarization error correction method for the microwave radiometer are solved. A new correction method is provided, based on the calibration of measured values of the microwave radiometer and the correction of the cold air overflow, multiple linear regression of the mode bright temperature data and the measured bright temperature of the microwave radiometer is carried outto directly obtain a correction matrix for all the cross polarization effects, and the cross polarization error correction of the measured bright temperature is realized. Compared with the existing cross polarization error correction method, the cross polarization error correction method has the advantages that correction can be completed without obtaining antenna pattern data, antenna installation deviation, satellite attitude and other parameters, and cross polarization errors from different sources can be uniformly corrected to improve the accuracy and speed of data processing.