109 results about "Standard Chinese" patented technology

The invention relates to a Chinese-English bilingual speech recognition method based on phoneme confusion. The method comprises the following steps: unifying a Chinese-English phoneme set by adopting a twice phoneme clustering method, obtaining a Chinese-English mixed acoustic model by retraining, correcting corresponding bilingual pronunciation dictionaries, and realizing Chinese-English bilingual recognition by a decoder based on the preceding steps. The Chinese recognition rate of a Chinese-English bilingual speech recognition system based on TCM phoneme confusion can be comparable to that of an independent Chinese speech recognition system; on the premise that English data with Chinese accent is unavailable and standard Chinese-English training data are used only, compared with the relatively independent English speech recognition system, recognition of English fragments with the Chinese accent is obviously improved; meanwhile, the Chinese-English bilingual speech recognition system based on the TCM phoneme confusion also has a better recognition performance than the existing common bilingual recognition system which performs phoneme clustering by virtue of a logarithm likelihood criterion, and has very high practicability.

The invention belongs to the technical field of mechanical arm control, and particularly relates to a method of writing standard Chinese characters by using a six-degree-of-freedom mechanical arm under the control of force. The method comprises the steps of extracting the contour information of Chinese characters, planning the path of the tail end of the mechanical arm, designing an impedance control method, finding the position of a contact surface and controlling the force of writing processes and particularly comprises the following steps of: extracting the position information of contour points of Chinese characters based on the TTF character library of Windows, and falling contour curves into straight lines and spline curves; and converting the extracted contour position coordinates of Chinese characters into the coordinates of Cartesian space of an end effector of the mechanical arm. The method has the beneficial effects that the six-degree-of-freedom mechanical arm is used for completing Chinese character writing tasks under the control of force; the method is universal for industrial mechanical arms, can be widely applied to practical teaching and science and technology exhibitions and can also be used for completing tasks such as spraying and engraving in industrial fields; Chinese characters can be written on brittle objects by using the method; and the accuracy and safety of writing are ensured.

The invention discloses a character library rapid generating method based on a characteristic template. The character library rapid generating method based on the characteristic template specificallycomprises the following steps that a set of standard Chinese character templates are collected, a user inputs specific sample words, and a user writing frame point set is collected; the writing characteristics of the user is calculated, and a standard template is modified according to the writing characteristics of the user to form a characteristic template; and the user written Chinese characteris splitted, and user writing parts are extracted; a whole-word frame point set is generated according to the characteristic template, and according to the writing pressure/speed of the user is subjected to Chinese character rendering, and the Chinese character is put into the library to generate a vector font. According to the method, the frame point set written by the user is mapped with the font library characteristic template, the character library is formed, and the making difficulty of the font library is greatly reduced, so that the manufacturing time of the font library is shortened, and the manufacturing cost of the font library is saved.

The invention discloses a Chinese character copying and writing method for a student tablet computer. The method includes the steps that standard Chinese characters are rasterized, the rasterized Chinese character data are uploaded to the computer to be stored, a simulated Chinese character writing model is called from a memory by the computer, all dot matrix information of a stroke is read and simulated and written according to the sequence of the strokes, then the middle line of the stroke dot matrix is computed, points are extracted at equal intervals on the middle line to serve as key points, all the key points of the stroke are extracted and are prompted remarkably with special figures, a user can conduct copying and writing following the track of the stroke prompted by the key point figures, and the computer recognizes the key point scoring condition automatically according to the prestored rasterized stand Chinese character data. By the adoption of the method, the Chinese character copying and writing method is easy to learn, vivid and capable of helping students master the stroke writing sequences, standard strokes and structures of the Chinese characters quickly and precisely.

The invention discloses a Chinese-Tibetan cross-language voice conversion method, and the method comprises the following steps: A, carrying out the cutting and marking, element classification and catalogue indexing of a voice corpus through designing a corresponding text corpus and recording the voice corpus, and completing the building of a Tibetan Lhasa dialect syllable database and a standard Chinese prosodic feature analysis database; B, building a fundamental frequency model through employing a five-degree tone model, building a duration conversion model and a pause duration conversion model and completing the building of a rhythm model; C, inputting a Tibetan text, selecting proper syllables from the built Tibetan Lhasa dialect syllable database through employing a decision tree, and completing the voice conversion based on the technology of waveform splicing technology; D, carrying out the modification of rhythm parameters (fundamental frequency, duration and pause duration) of the converted voice through employing a STRAIGHT algorithm, completing the rhythm control, and outputting standard Chinese voices. The method achieves a Chinese-Tibetan cross-language voice conversion system based on rhythm control.

The invention discloses a method used for identifying isolated word of standard Chinese on basis of basic frequency envelope. The method comprises: a vocabulary is stored in a template library in advance in the form of basic frequency envelope. The method further comprising the steps in sequence as follows: 1) the speech is cut and the noise is reduced; 2) the basic frequency characteristics of the speech are extracted; 3) judgment: when the language material is used for training, that is to say, when no corresponding vocabulary exists in a corpus, the step 4) is started; when the language material is used for identifying, the step 5) is started; 4) the basic frequency characteristics of the language material is added into the template library; 5) the similarities between the test speech and the basic frequency envelope of the template are compared and the vocabulary represented by the template with the minimum distance to the test speech is taken as the identification result; 6) the identification result is output. By researching the tone of Chinese, the method of the invention is capable of identifying the isolated word of standard Chinese by comparing the similarities between the test speech and the basic frequency envelope of the template.

The invention discloses a Chinese sign language two-way translation system, method and device. The Chinese sign language two-way translation system comprises a sign language model base module, an image identifying module, a sign language model matching module and a forward output module. The sign language model base module is used for storing the mathematical models of standard Chinese sign languages. The mathematical model of each sign language corresponds to identification information. The image identifying module is used for identifying a human body part and the movement information of the human body part to generate a gesture model according to the human body part and the corresponding movement information. The sign language model matching module is used for acquiring identification information corresponding to a mathematical model with the matching degree higher than a preset value. The forward output module is used for outputting corresponding forward output information according to the identification information. According to the invention, a human body part and the movement information of the human body part are identified; a gesture model is generated based on the human body part and the corresponding movement information; and the gesture identification accuracy is improved through a data fusion technology.

The invention discloses a method for evaluating attractiveness of handwritten Chinese characters. The method includes the following specific implementation steps that a bounding box mesh structure of the Chinese characters is established; stroke segments included in standard Chinese characters are judged as ones of horizontal strokes, vertical strokes, left-descending strokes and right-descending strokes, feature points of the Chinese characters are extracted; the feature points of the standard Chinese characters are mapped to corresponding meshes, bordering processing and relaxation calculation are conducted on the meshes, and then the deformation meshes, in a bounding box, of the standard Chinese characters are acquired; pixel point sets of the handwritten Chinese characters are acquired, the stroke segments are separated and judged, and then the feature points of the handwritten Chinese characters are extracted; secondary deformation meshes, formed on the basis of the deformation meshes of the standard Chinese characters, of the handwritten Chinese characters are acquired; energy differences between the two deformation meshes are calculated; attractiveness of the user handwritten Chinese characters is evaluated according to the energy differences. According to the method for evaluating attractiveness of the handwritten Chinese characters, the problem that the handwritten Chinese characters are easy to implement but difficult to evaluate in the exercise process can be effectively solved, and the method has important application value.

Methods and systems for romanizing Chinese ideograms allow a user to create a personalized spelling dictionary that converts a user's desired roman-alphabet spelling to an equivalent Chinese character. A phonetic combination from a standard Chinese dialect is selected. The user defines a roman alphabet equivalent of the selected phonetic combination that fits the way the user pronounces the phonetic combination in the user's own dialect or idiolect.

The invention discloses a recognition method and device for Chinese character writing. The recognition method and device are used for recognizing a Chinese character writing process. The recognition method comprises the steps that the standard stoke data of standard Chinese characters are pre-stored, wherein the standard stoke data include the information of the sequence of the standard strokes of the standard Chinese characters and the information of the directions and positions of the standard strokes; the real-time stoke data of a real-time Chinese character input by a user are collected, wherein the real-time Chinese character is the same as a specific Chinese character, the specific Chinese character is one of the standard Chinese characters, and the real-time stoke data include the information of the sequence of the real-time strokes of the real-time Chinese character and the information of the directions and positions of the real-time strokes; the corresponding relation between the standard stokes of the specific Chinese character and the real-time stokes of the real-time Chinese character is obtained according to the information of the directions and the positions of the real-time strokes and the information of the directions and the positions of the standard stokes of the specific Chinese character; the standard stoke data of the specific Chinese character are compared with the real-time stoke data of the real-time Chinese character according to the corresponding relation, and a recognition result of the correctness of the real-time Chinese character writing process is obtained.

The invention discloses a handwritten model training method, a handwriting recognition method, a device, equipment and a medium. The handwritten model training method comprises: acquiring a standard Chinese character training sample, initializing a convolutional neural network, inputting the standard Chinese character training sample into the convolutional neural network for training, updating theweight and offset of the neural network by using a backward propagation algorithm based on stochastic gradient descent, and obtaining the standard Chinese character recognition model; obtaining and adopting the non-standard Chinese character training sample, training, obtaining and adjusting the Chinese handwriting recognition model; obtaining and adopting the Chinese character sample to be tested to obtain the error word training sample; and using the erroneous word training sample based on the backward gradient propagation algorithm of batch gradient descent to update the weight and offsetof the Chinese handwriting recognition model so as to obtain the target Chinese handwriting recognition model. By adopting the handwritten model training method, a target Chinese handwriting recognition model with high recognition rate of handwriting may be obtained.

The invention relates to a voice data mapping method and apparatus. The voice data mapping method comprises: acquiring voice pronunciation data and establishing a local native language voice database, a local Chinese voice database, and a standard Chinese voice database; establishing a scoring voice database; performing analysis and statistic on the scoring voice database in order to acquire typical biased errors of local Chinese pronunciation; analyzing voice data in the local native language voice database, the local Chinese voice database, and the scoring voice database in order to obtain expert knowledge data and establishing an expert knowledge database; selecting a characteristic vector according to the expert knowledge data in the expert knowledge database and extracting voice parameters from the expert knowledge data, wherein the characteristic vector comprises characteristic parameters and the weight of the characteristic parameters; and establishing a scoring mapping relation between the characteristic vector and scoring data. The voice data mapping method and apparatus is capable of establishing an accurate scoring mapping relation between the characteristic vectors and scoring data based on language ecology.

The invention relates to a method for extracting author affiliation information of English literature published by Chinese authors. The method is used for extracting the Chinese name information of the affiliations of the Chinese authors from an English literature library and includes: using web crawler to acquire the bibliographic data of all related English papers published by the Chinese authors from the English literature library; extracting paper titles, author affiliation information and publishing time from the bibliographic data; processing the author affiliation information to allow the author affiliation information to correspond to the standard Chinese names of author affiliations; saving the extracted paper titles, author affiliation information, publishing time and the standard Chinese names of the affiliations into a self-built database for follow-up inquiry and statistical counting. By the method, search result accuracy is guaranteed to a large degree, the process of manual affiliation information inquiring and checking is avoided, a user can inquire and statistically count the English literature information published by the affiliations, and high recall ratio and accuracy are achieved.

The invention discloses a stroke addition recognition method used when Chinese characters are handwritten on a digital handwriting platform in an online mode. The method specifically comprises the steps that a standard Chinese character stroke type code library identified with a direction code set is established; a tracing point set generated when a user handwrites Chinese characters in the online mode is recorded and is processed, and a record file of stroke type codes identified with the direction code set by the tracing point set is established; the stroke types of a current stroke and a prior stroke and the positional relation between the current stroke and the prior stroke are judged by combining traversing comparison conducted between the current stroke and the prior stroke with comparison conducted between the current stroke and the corresponding Chinese character stroke type code in the standard Chinese character stroke type code library, so that whether a stroke needs to be added is judged, and the point sets of the two strokes in the addition relation are combined to enable a complete stroke to be obtained. By the adoption of the stroke addition recognition method for the online handwritten Chinese characters, the stroke addition behavior which possibly exists when the user handwrites the Chinese characters in the online mode can be recognized, and the method has great significance in evaluating Chinese character writing quality, guiding Chinese character writing and recognizing the Chinese characters.

The invention provides a point guide calligraphy learning device and a manufacturing method thereof, including the following procedures that: (1) a calligraphy learning carrier comprising a writing plane is set; (2) more than one character frame is arranged on the writing plane; (3) a character frame is provided inside with a reference Chinese character; (4) other character frames are provided inside with point Chinese characters. The calligraphy learning carrier comprises a handwriting board, a display screen board and paper. The point guide calligraphy learning device of the invention can let a Chinese character practicer enjoy the gradual practicing process by changing the number of 'points' of each character, the distance between 'points' of each character and the depth of writing, thereby the practicer can write order and standard Chinese characters; moreover, the invention also leaves further development space for the practicer in order to form the writing characteristic and style of the practicer.

The invention discloses a Chinese geographic coding and decoding method. The method comprises the following steps: a GIS (Geographic Information System) including an electronic map, coordinate data and geological information data is called for use, wherein each map unit on the electronic map corresponds to one coordinate data and a group of geographic information data; as for each map unit, a standard Chinese geographic indication is set and stored in a standard data base, the storage position of the standard Chinese geographic indication in the standard data base is mapped in the geographic information data, and the standard Chinese geographic indication comprises separator fields, pointing fields and identification fields; and as for each standard Chinese geographic indication, a fuzzy Chinese geographic indication is set and stored in a fuzzy data base, the storage position of the fuzzy Chinese geographic indication in the fuzzy data base is mapped in the geographic information data, the mapping of the fuzzy Chinese geographic indication and that of the standard Chinese geographic indication are correlated, and the fuzzy Chinese geographic indication comprises a primary fuzzy Chinese geographic indication and a secondary fuzzy Chinese geographic indication.

The invention discloses a broadcast signal prejudging and preprocessing method. The method includes the steps that a speech segment is intercepted, and the segment with the speech and the segment without the speech are separated; a speech time domain sequence containing the segment with the speech is subjected to framing and sliding window adding to obtain a speech time domain sequence, the additive noise is removed, and a speech time domain sequence after additive noise removal is obtained; the speech time domain sequence is filtered to obtain a speech time domain sequence after impulse noise removal; a recirculating network neural model is established, the segment with the speech after impulse noise removal is screened, and a first speech segment containing standard Chinese speeches only and a second speech segment containing both the standard Chinese speeches and the background music are screened out; and the background music of the second speech segment is separated from the standard Chinese speeches. A pure broadcast recording file for eliminating the current noise, the ambient sound, speaker modal particle and the background music is obtained, thus the accuracy of results obtained by keyword recognition is improved, and the mis-recognition rate is decreased.

The invention discloses a handwriting input processing method and a terminal device. The method comprises the steps that a first handwriting input by a user is obtained, and the first handwriting is matched with the user personal handwriting in a database; if the first handwriting is matched with the user personal handwriting, the user personal handwriting matched with the first handwriting is obtained, then a corresponding standard Chinese character is obtained according to the obtained user personal handwriting, and the obtained standard Chinese character is used as a preferable character to be fed back to the user. In this way, the handwriting character input by handwriting of the user can be rapidly and accurately recognized, and the handwriting input efficiency is improved.

The invention discloses a handwriting model training method, a text recognition method, a device, an apparatus and a medium. A method for training a handwriting model includes: a Chinese text train sample is obtained, Chinese character train samples in that specification are batch divided according to preset batches, the batched Chinese text train samples are inputted into a circulating neural network, and trained based on a continuous-time classification algorithm, and network parameters are updated by a time-dependent back propagation algorithm to obtain a Chinese text recognition model of the specification; non-standard Chinese text training samples are obtained and used for training to obtain adjusted Chinese handwritten text recognition model; Chinese text sample to be tested are acquired and used to obtain error text training samples; the network parameters of Chinese handwritten text recognition model are updated by training samples of error text, and the target Chinese handwritten text recognition model is obtained. A target Chinese handwritten text recognition model with high recognition rate can be obtain by using that handwritten model train method.

The invention discloses a Chinese character writing font identification method and a Chinese character writing font identification device which are used for identifying the font of a Chinese character. The method comprises the following steps: storing standard stroke data of a standard Chinese character, wherein the standard stroke data includes the coordinates of the points constituting standard strokes; acquiring first real-time stroke data of a real-time Chinese character input by a user in a stroke way, wherein the first real-time stroke data includes the coordinates of the points constituting real-time strokes; calculating the center of gravity of the standard Chinese character according to the standard stroke data, and calculating the center of gravity of the real-time Chinese character according to the first real-time stroke data; shifting the real-time Chinese character by the difference between the center of gravity of the real-time Chinese character and the center of gravity of the standard Chinese character, and acquiring second real-time stroke data of the shifted real-time Chinese character, wherein the second real-time stroke data includes the coordinates of the points constituting the shifted real-time strokes; and calculating the distance between each shifted real-time stroke and the corresponding standard stroke according to the coordinates of the points constituting the shifted real-time stroke and the coordinates of the points constituting the standard stroke.

The invention discloses a Chinese character skeleton stroke segment processing method and system. The method comprises the following steps: obtaining all skeletal stroke segments of each Chinese character; determining an average width value of all skeletal pen segments; judging the type of the skeleton pen segment according to the adjacency attribute and the average width value of each skeleton pen segment, wherein the type of the skeleton pen segment is a burr type, an internal short pen segment type and a long pen segment type; according to the type of the skeletal stroke segment, the skeletal stroke segment is respectively disambiguated to obtain the skeletal stroke segment of the Chinese character corresponding to the standard Chinese character. The invention solves the technical problems existing in the prior art in the process of disambiguation and sorting of Chinese character stroke segments.

Disclosed is a Chinese characters input method with phrase association and speech prompt in mobile information terminals. The method comprises: inputting standard Chinese characters by combining the standard phoneticize input method and TTS speech association method, selecting standard phoneticize input method when inputting information, completing standard phoneticize input via corresponding letters on press-keys, and the single arrayed characters are covered by input items, arranging the pronunciations of phrases interpretations in corresponding database, and outputting arranged pronunciation interpretations with corresponding speech, in the phoneticize mode, making user select right Chinese characters input via phrase association interpretation and speech prompt.

The invention discloses a method for recognizing the writing of a Chinese character. The method comprises the steps of (1) establishing a standard set which includes a stroke order, a stroke type anda stroke number of each standard Chinese character, (2) collecting a feature set of a handwritten input Chinese character, wherein the feature set includes a stroke order, a stroke type and a stroke number of the handwritten input Chinese character, and (3) comparing the feature set and the standard set through a comparison device and obtaining a correctness recognition result of the handwritten input Chinese character according to a comparison of the feature set and the standard set. Through comparing the standard set with the collection of Chinese character writings and the feature set of awriting period, the stroke order of a writer in a writing process can be effectively compared, thus a problem in the writing period can be judged, the writer can be effectively helped to change writing habits, and errors in the writing period are corrected.

A processing method of a snack food of Chinese chestnut is provided. The processing method comprises the following steps: taking fresh Chinese chestnut from Qianxi, grading and screening fresh Chinese chestnut fruits, and singling out fruits with wormholes, mildewed and rotten fruits, and impurities; manually cutting surfaces of the fresh Chinese chestnut fruits, stir-baking the cut fresh Chinese chestnut fruits in a wok, removing shells and endothelia of the Chinese chestnut fruits, and impurities, singling out non-standard Chinese chestnut kernels, carrying out quality inspection, rinsing with clear water, removing the shells and the endothelia of the Chinese chestnut fruits, and impurities, and draining remained Chinese chestnut kernels; quick-freezing the drained Chinese chestnut kernels in a -38DEG C cold storage; and roasting or frying the well quick-frozen Chinese chestnut kernels to prepare the snack food of Chinese chestnut. Above quick-freezing manner adopted in the invention makes nutritional components of Chinese chestnut be effectively reserved, and the processing of Chinese chestnut to the snack food has the advantages of simple processing method, no loss of the nutritional components, delicacy, and improvement of the grade and the nutritional components of Chinese chestnut. So a new edible use of Chinese chestnut is developed.

The invention discloses an objective examination method of breach sound and affricate in standard Chinese, including the steps: receiving input voice; syncopating the input voice; distilling distinguishing characteristics; giving a mark according to an evaluating model; and obtaining a pronunciation score. By applying the objective examination method and adopting the distinguishing characteristics which can better reflect the pronunciation essential to distinguish the breach sound and the affricate, the better distinguishing performance can be obtained.

The invention discloses a handwriting model training method, a handwriting character recognition method and device, an apparatus and a medium. The handwriting model training method comprises the following steps of: obtaining a standard Chinese character training sample, and batching the standard Chinese character training sample according to a preset batch; obtaining the Chinese character recognition model by updating the network parameters with time-dependent back propagation algorithm and training the batched Chinese character training samples with bi-directional long-short time memory neural network; obtaining and using non-standard Chinese character training samples, and training to obtain the adjusted Chinese handwritten character recognition model; obtaining a Chinese character sample to be tested and adopting the same to obtain an error character training sample; based on the time-dependent back propagation algorithm of batch gradient descent, updating the network parameters ofChinese handwritten character recognition model by training samples of error characters, and obtaining the target Chinese handwritten character recognition model. A target Chinese handwritten character recognition model with high recognition rate can be obtained by using that handwritten model train method.