CN105404621A - Method and system for blind people to read Chinese character - Google Patents
Method and system for blind people to read Chinese character Download PDFInfo
- Publication number
- CN105404621A CN105404621A CN201510623525.5A CN201510623525A CN105404621A CN 105404621 A CN105404621 A CN 105404621A CN 201510623525 A CN201510623525 A CN 201510623525A CN 105404621 A CN105404621 A CN 105404621A
- Authority
- CN
- China
- Prior art keywords
- braille
- word
- participle
- polyphone
- chinese character
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Document Processing Apparatus (AREA)
Abstract
The invention provides a method and a system for blind people to read Chinese characters, and relates to the technical field of natural language processing and the technical field of disabled-oriented human-computer interaction. The method comprises the following steps: obtaining a Chinese language text, carrying out a word segmentation operation on the Chinese language text to generate a Chinese character string, converting each word in the Chinese character string into corresponding Pinyin by referring to part-of-speech tagging obtained by word segmentation through a pronouncing dictionary, a polyphone dictionary and word frequency information, and connecting the Pinyin into a Pinyin string; looking up a Pinyin and blind character contrast dictionary, converting the Pinyin string into a blind character string, carrying out braille word segmentation on the blind character string through a word segmentation model to generate initial braille segmentation words, fusing the Chinese character string with the initial braille segmentation words to generate new braille segmentation words, and regulating the new braille segmentation words according to a braille segmentation word ligature rule; and carrying out braille tone marking on the new braille segmentation words regulated according to the braille segmentation word ligature rule to generate final braille segmentation words, and displaying the final braille segmentation words.
Description
Technical field
The present invention relates to natural language processing technique field and the human-computer interaction technique field towards disabled person, particularly relate to a kind of method and system reading Chinese character for blind person.
Background technology
In current information society, the level of IT application improves constantly, infotech is widely applied in the work, studying and living of people, and internet also becomes an important component part in people's daily life, network in a kind of mode easily for people provide the information resources of magnanimity.In China, various digitizing, network text resource store mainly with the form of Chinese language text greatly, and these resources to be difficult to use by existing 1,200 ten thousand blind persons of China.Which prevent the information resources that blind person enjoys magnanimity as normal person, the telecoms gap between blind person and normal person is constantly expanded, the survival and development ability of blind person in informationized society is subject to further restriction.Although existing speech synthesis technique reaches its maturity, textual resources a large amount of on network can be converted to audio file by phonetic synthesis and make blind person can obtain these information by the sense of hearing, but the storage of voice resource is compared and is expended space, and it is inconvenient carrying, in inquiry etc., and voice channel obtaining information efficiency is lower, therefore, for blind person, read text resource remains the most important mode of acquisition information.
The word that China blind person uses when reading writing is Chinese braille, Chinese braille is based on Blair (Braille) braille system, each blind symbol all using two row 6 totally o'clock as a basic structure, the projection that this 6 points have, what have is not protruding, form 64 kinds of changes, 64 kinds of different characters can be represented.In Chinese braille, often kind of character represents initial consonant, simple or compound vowel of a Chinese syllable or a tone in the Chinese phonetic alphabet respectively, and different characters is according to Chinese phonetic alphabet rule composition legal syllables to represent Chinese character, and therefore, Chinese braille is a kind of alphabetic writing in essence.Braille generally prints and writes on special thicker braille paper, braille paper extrudes protruding some position and touches for blind person and read.Touch on computers for enabling blind person and read braille, currently be equipped with and produced blind aobvious device, this equipment can be connected with computing machine, blind symbol string in receiving computer, and it is shown as corresponding protruding some position on the aobvious machine side plate of point, after receiving new blind symbol string, original some position can be removed again show new some position on panel.
Although there has been the aobvious device of point, blind person has still been difficult to read Chinese language text on computers, and reason is also to need Chinese language text to be converted to braille.Due to phenomenons such as the ubiquitous sound multiword of Chinese, a word multitone, make Chinese corresponding to the not simple rule of conversion of braille, and need to consider grammer, semanteme etc.What is more important, braille has word link writing rule, requires the word or phrase that possess certain semanteme to separate with one " short side ", so that blind person understands.Current existing method generally adjusts the braille after obtaining participle based on braille word link writing rule to Chinese word segmenting result, but because braille word link writing rule is general relevant to semanteme and have certain subjectivity, therefore, when automatically being completed by computing machine, participle accuracy rate is lower, after these methods of use are changed, also need to do a large amount of artificial correction work, cause inefficiency, also make that the time of the acquisition of braille textual resources is longer and cost is higher.Therefore, improve the accuracy rate of the blind conversion of the Chinese, reduce the artificial operation revised, accelerate the efficiency of the blind conversion of the Chinese, for the popularity rate of raising Chinese information resource in blind community, allowing blind community incorporate better in mainstream society has important realistic meaning.
Summary of the invention
For the deficiencies in the prior art, the present invention proposes a kind of method and system reading Chinese character for blind person.
The present invention proposes a kind of method reading Chinese character for blind person, comprising:
Step 1, obtains Chinese language text, carries out participle operation to described Chinese language text, generate Chinese character string, by pronunciation dictionary, polyphone dictionary and word frequency information, with reference to the part-of-speech tagging that participle obtains, each word in described Chinese character string be converted to corresponding phonetic and connect for pinyin string;
Step 2, by searching the contrast dictionary of phonetic and blind symbol, described pinyin string is converted to blind symbol string, by participle model, braille participle is carried out to described blind symbol string, generate initial braille participle, described Chinese character string and described initial braille participle are merged, generates new braille participle, according to braille word link writing rule, described new braille participle is adjusted;
Step 3, carries out braille mark tune to according to the described new braille participle after braille word link writing rule adjustment, generates final braille participle, is shown by described final braille participle.
The described method reading Chinese character for blind person, the concrete steps in described step 1, described Chinese character string being converted to pinyin string are:
Step 2.1, for each word in described Chinese character string, judges whether each word is multi-character words, if multi-character words, and the phonetic that described multi-character words can be found in pronunciation dictionary corresponding, then directly return phonetic corresponding to described multi-character words, otherwise perform step 2.2;
Described multi-character words cutting is the sequence of Chinese character by step 2.2, gets Chinese characters all in described multi-character words successively, to each Chinese character, performs step 2.3 to 2.4;
Step 2.3, for current Chinese character, searches polyphone dictionary, judges whether described current Chinese character is polyphone, if not polyphone, searches the phonetic of described current Chinese character and return described phonetic in pronunciation dictionary; Otherwise perform step 2.4;
If step 2.4 polyphone, then perform following steps, concrete steps are:
If the current polyphone of step 2.4.1 comes from a monosyllabic word, then directly perform step 2.4.2; If multi-character words, then perform following step:
For the polyphone w in multi-character words
k, a) step, forms the word W of a n+1 word with a follow-up n word
k,n=w
kw
k+1w
k+n, in polyphone phrase dictionary, search W
k,n, as found, then with W
k,nin be searched the pronunciation of word as polyphone w
kpronunciation and return; As do not found, then performing b) step, forming the word W of a n+1 word with n word above
n-k, k=w
n-kw
n-kk+1w
n, in polyphone phrase dictionary, search W
n-k, k, as found, then with W
k,nin be searched word pronunciation as polyphone pronunciation and return, as do not searched, then respectively with follow-up and above n-1 word form the word W of a n word
k, n-1, W
n-k+1, k, described multi-character words to be performed a) respectively, b) step, until determine described polyphone w
kpronunciation;
Step 2.4.2 supposes that described polyphone has tone
1..., tone
nn pronunciation altogether, participle part of speech definition of probability is P
pos, weights are λ
1, probabilistic language model is defined as P
lm, weights are λ
2, participle word frequency definition of probability is P
freq, weights are λ
3, system is that each pronunciation of described polyphone calculates a score Score
i, wherein Score
i=λ
1p
pos(tone
i)+λ
2p
lm(tone
i)+λ
3p
freq(tone
i), take out the pronunciation that score is the highest
as polyphone final phonetic and return.
The described method reading Chinese character for blind person, the step of carrying out in described step 2 merging is, for described Chinese character string C=c
1c
2c
mwith described initial braille participle B=b
1b
2b
n, wherein c
i, b
jrepresent a participle in described Chinese character string and described initial braille participle respectively, for described initial braille participle B, B is mapped to corresponding described Chinese character string B'=b
1' b'
2b'
n, wherein b'
jfor described initial braille participle b
jbe mapped as the participle after Chinese.
The described method reading Chinese character for blind person, in described step 2, braille word link writing rule is as follows:
Combination handwriting rule: POS
k: [m, n]: POS
k-m+ ... + POS
k+ ... + POS
k+n→ POS
k-mpOS
k+n
Word segmentation regulation:
POS
kfor activation condition, m and n represents front m the word and n word that need to check current new braille participle respectively, if m and n is 0, then represent that this is a word segmentation regulation, represent after second colon be participle part of speech combination, if meet this combination, then perform the operation after right arrow.
The described method reading Chinese character for blind person, the concrete steps that braille mark described in described step 3 is adjusted are:
Check the phonetic of the corresponding word of the described new braille participle after each adjustment successively, and the rule in assembling with braille mark is compared, if satisfied condition, then carry out mark to current new braille participle and adjust, the form that described braille mark is assembled is as follows:
Mark adjusts rule: tone
k: [n]: tone
ktone
k+n
Wherein tone
kfor the phonetic of current new braille participle, n is the phonetic needing to check rear n of current new braille participle new braille participle, tone
ktone
k+nfor mark tune condition, if pinyin sequence meets mark tune condition, then to tone
kcarry out mark to adjust.
The present invention also proposes a kind of system reading Chinese character for blind person, comprising:
Obtain pinyin string module, for obtaining Chinese language text, participle operation is carried out to described Chinese language text, generate Chinese character string, by pronunciation dictionary, polyphone dictionary and word frequency information, with reference to the part-of-speech tagging that participle obtains, each word in described Chinese character string be converted to corresponding phonetic and connect for pinyin string;
Obtain new braille participle and adjusting module, for the contrast dictionary by searching phonetic and blind symbol, described pinyin string is converted to blind symbol string, by participle model, braille participle is carried out to described blind symbol string, generate initial braille participle, described Chinese character string and described initial braille participle are merged, generates new braille participle, according to braille word link writing rule, described new braille participle is adjusted;
Braille display module, for carrying out braille mark tune to according to the described new braille participle after braille word link writing rule adjustment, generates final braille participle, is shown by described final braille participle.
The described system reading Chinese character for blind person, the concrete steps in described acquisition pinyin string module, described Chinese character string being converted to pinyin string are:
Step 2.1, for each word in described Chinese character string, judges whether each word is multi-character words, if multi-character words, and the phonetic that described multi-character words can be found in pronunciation dictionary corresponding, then directly return phonetic corresponding to described multi-character words, otherwise perform step 2.2;
Described multi-character words cutting is the sequence of Chinese character by step 2.2, gets Chinese characters all in described multi-character words successively, to each Chinese character, performs step 2.3 to 2.4;
Step 2.3, for current Chinese character, searches polyphone dictionary, judges whether described current Chinese character is polyphone, if not polyphone, searches the phonetic of described current Chinese character and return described phonetic in pronunciation dictionary; Otherwise perform step 2.4;
If step 2.4 polyphone, then perform following steps, concrete steps are:
If the current polyphone of step 2.4.1 comes from a monosyllabic word, then directly perform step 2.4.2; If multi-character words, then perform following step:
For the polyphone w in multi-character words
k, a) step, forms the word W of a n+1 word with a follow-up n word
k,n=w
kw
k+1w
k+n, in polyphone phrase dictionary, search W
k,n, as found, then with W
k,nin be searched the pronunciation of word as polyphone w
kpronunciation and return; As do not found, then performing b) step, forming the word W of a n+1 word with n word above
n-k, k=w
n-kw
n-kk+1w
n, in polyphone phrase dictionary, search W
n-k, k, as found, then with W
k,nin be searched word pronunciation as polyphone pronunciation and return, as do not searched, then respectively with follow-up and above n-1 word form the word W of a n word
k, n-1, W
n-k+1, k, described multi-character words to be performed a) respectively, b) step, until determine described polyphone w
kpronunciation;
Step 2.4.2 supposes that described polyphone has tone
1..., tone
nn pronunciation altogether, participle part of speech definition of probability is P
pos, weights are λ
1, probabilistic language model is defined as P
lm, weights are λ
2, participle word frequency definition of probability is P
freq, weights are λ
3, system is that each pronunciation of described polyphone calculates a score Score
i, wherein Score
i=λ
1p
pos(tone
i)+λ
2p
lm(tone
i)+λ
3p
freq(tone
i), take out the pronunciation that score is the highest
as polyphone final phonetic and return.
The described system reading Chinese character for blind person, the new braille participle of described acquisition the step of carrying out in adjusting module merging are, for described Chinese character string C=c
1c
2c
mwith described initial braille participle B=b
1b
2b
n, wherein c
i, b
jrepresent a participle in described Chinese character string and described initial braille participle respectively, for described initial braille participle B, B is mapped to corresponding described Chinese character string B'=b
1' b'
2b'
n, wherein b'
jfor described initial braille participle b
jbe mapped as the participle after Chinese.
The described system reading Chinese character for blind person, described acquisition new braille participle and in adjusting module braille word link writing rule as follows:
Combination handwriting rule: POS
k: [m, n]: POS
k-m+ ... + POS
k+ ... + POS
k+n→ POS
k-mpOS
k+n
Word segmentation regulation:
POS
kfor activation condition, m and n represents front m the word and n word that need to check current new braille participle respectively, if m and n is 0, then represent that this is a word segmentation regulation, represent after second colon be participle part of speech combination, if meet this combination, then perform the operation after right arrow.
The described system reading Chinese character for blind person, the concrete steps that braille mark described in described braille display module is adjusted are:
Check the phonetic of the corresponding word of the described new braille participle after each adjustment successively, and the rule in assembling with braille mark is compared, if satisfied condition, then carry out mark to current new braille participle and adjust, the form that described braille mark is assembled is as follows:
Mark adjusts rule: tone
k: [n]: tone
ktone
k+n
Wherein tone
kfor the phonetic of current new braille participle, n is the phonetic needing to check rear n of current new braille participle new braille participle, tone
ktone
k+nfor mark tune condition, if pinyin sequence meets mark tune condition, then to tone
kcarry out mark to adjust.
From above scheme, the invention has the advantages that:
The present invention is different from the blind switch technology of the existing Chinese, first Chinese word segmenting is carried out to Chinese character string, in word segmentation result, use the word link writing of series of complex rule to carry out the way of secondary treating again, the present invention utilizes the braille participle model of the Corpus--based Method machine learning techniques of structure directly to carry out single step participle to blind symbol string, word segmentation result meets braille word link writing rule substantially, only need carry out a small amount of fine setting and namely can be used as braille output, compared to existing technology, avoid by computer disposal complexity, relate to the problem that the regular accuracy rate caused of semantic word link writing is not high, participle accuracy rate and the blind conversion accuracy of the overall Chinese have larger lifting.
Accompanying drawing explanation
Fig. 1 is the method flow diagram reading Chinese character for blind person;
Fig. 2 is the process flow diagram that Chinese character string after participle is converted to pinyin string.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly, below in conjunction with drawings and Examples, the method reading Chinese character for blind person of the present invention is further elaborated, be to be understood that, specific embodiment described herein only in order to explain the present invention, is not intended to limit the present invention.
Of the present inventionly read the method main flow of Chinese character as shown in Figure 1 for blind person, it is input as a Chinese sentence, i.e. a Chinese character string, exports as corresponding braille, and is presented on blind aobvious device.
Step 1. Chinese word segmenting.Namely Chinese word segmentation system is adopted to be the sequence of Chinese word by the Chinese character string cutting of input, obtain the Chinese character string after participle, it is each word mark part of speech simultaneously, Chinese word segmenting can adopt current existing various method and system, as based on the maximum of dictionary or smallest match method, based on the method for Hidden Markov Model (HMM) (HMM), based on the method etc. of maximum entropy model;
Chinese character string after participle is converted to pinyin string by step 2., namely pronunciation dictionary, polyphone dictionary and word frequency information is utilized, with reference to the part-of-speech tagging that participle obtains, each word in Chinese character string after participle be converted to corresponding phonetic and connect for pinyin string, described pronunciation dictionary is the mapping table of Chinese words (comprising monosyllabic word and multi-character words) and phonetic.In one embodiment, the scale of pronunciation dictionary is about 70,000 words, all polyphones and multiple phonetics corresponding to each polyphone thereof are listed in described polyphone dictionary, described word frequency information is the frequency of occurrences in Chinese language text of each Chinese character, and this information adopts a large amount of Chinese language text to add up in advance and obtains.In one embodiment, in word frequency information, the scale of word is about 7000 words.
Be below the concrete steps of participle, as shown in Figure 2:
Step 2.1 is for each word in the Chinese character string after participle, judge whether this word is multi-character words (comprising two or more Chinese characters), if multi-character words, and can find the phonetic that this word is corresponding in pronunciation dictionary, then directly return this phonetic, otherwise perform step 2.2;
Word segmentation, for the word (monosyllabic word or multi-character words) of input, is the sequence of Chinese character, gets the Chinese character that they are all successively by step 2.2, to each Chinese character, performs step 2.3 to 2.4;
Step 2.3, for current Chinese character, searches polyphone dictionary, judges whether this word is polyphone, if not polyphone, searches the phonetic of this word and return this phonetic in pronunciation dictionary; Otherwise perform step 2.4;
Step 2.4, for polyphone, needs the phonetic of comprehensive much information determination polyphone.Concrete steps are:
If the current polyphone of step 2.4.1 comes from a monosyllabic word, then directly perform step 2.4.2; Otherwise first perform following step:
For the polyphone w in multi-character words
k, the word W of a n+1 word a) is formed with a follow-up n word
k,n=w
kw
k+1w
k+n, in polyphone phrase dictionary, search W
k,n, as found, then return using the pronunciation of this word in this phrase as the pronunciation of polyphone; As do not found, then b) form the word W of a n+1 word with n word above
n-k, k=w
n-kw
n-kk+1w
n, in polyphone phrase dictionary, search W
n-k, k, as found, then return using the pronunciation of this word in this phrase as the pronunciation of polyphone, as do not searched, then respectively with follow-up and above n-1 word form the word W of a n word
k, n-1, W
n-k+1, k, this word to be performed a) respectively, b) step, until determine that this polyphone pronounces.If during n=1, W
k, k+1, W
k-1, kstill at polyphone phrase dictionary lookup to pronunciation, then sky cannot be returned;
Step 2.4.2 supposes that polyphone has tone
1..., tone
nn pronunciation altogether, participle part of speech definition of probability is P
pos, weights are λ
1, probabilistic language model is defined as P
lm, weights are λ
2, participle word frequency definition of probability is P
freq, weights are λ
3, system is that each pronunciation of polyphone calculates a score Score
i, wherein Score
i=λ
1p
pos(tone
i)+λ
2p
lm(tone
i)+λ
3p
freq(tone
i), take out the pronunciation that score is the highest
as polyphone final phonetic and return.It should be noted that, for the probability of part of speech, word frequency, language model each pronunciation all types of, need to be normalized, all types of weights can set based on experience value.
Pinyin string is converted to blind symbol string by step 3..By searching the contrast dictionary of phonetic and blind symbol, pinyin string step 2 obtained is converted to blind symbol string, and blind symbol string is now the blind symbol string of non-participle.The contrast dictionary of described phonetic and blind symbol is the mapping table of phonetic and the blind symbol of correspondence.
Step 4. adopts the participle model trained by statistical machine learning method in advance to carry out braille participle, generates initial braille participle.Adopt the perceptron model that current association area is conventional, adopt the braille language material having divided word during model training, the feature of employing is unitary feature, binary feature and attributive character.During participle to blind symbol string each can cutting position extract feature also utilize the model calculating probability trained, carry out word segmentation according to probabilistic determination the need of in this position.
Training pattern adopts perceptron algorithm, and learn from the discriminant mapping model being input to output, input is the sentence in corpus, and output is corresponding annotation results.
Word disaggregated model is adopted to the participle of braille sentence.A given sentence be made up of n word, the process of participle is that this sentence is divided into m (m≤n) block, and each block is a significant word.Distribute the class mark that represents its position in word to each word, participle problem is changed into word classification problem.Adopt b, m, e, s as the border class mark of word, b, m, e represent this word respectively and are positioned at the starting position of word, centre position, end position, and this word of behalf is monosyllabic word.Decode procedure finds the annotated sequence y making goals for evaluation function f (x) the highest.
Wherein, f (x) mark is accumulated each word and the right mark of class mark, (i, t) ∈ y (s.t.1≤i≤n, t ∈ b, m, e, s}), Φ (x, y) is feature extraction function,
it is parameter vector.Participle uses Viterbi decoding algorithm.
Step 5. Chinese and initial braille participle merge, and namely utilize Chinese braille word segmentation result to finely tune braille word segmentation result, to improve the accuracy rate of participle further.
For Chinese word segmentation C=c
1c
2c
mwith braille participle B=b
1b
2b
n, wherein c
i, b
jrepresent a participle in Chinese and braille respectively, for braille participle B, B can be mapped to corresponding Chinese word segmentation B'=b
1' b'
2b'
n, wherein b'
jfor braille participle b
jbe mapped as the participle after Chinese.Editing distance alignment is carried out to Chinese word segmentation C and the braille participle B' being mapped as Chinese, fragments different in C and B' can be obtained, use above-mentioned fusion rule, determine that the net result of different fragments adopts Chinese word segmentation result or braille word segmentation result.Suppose that fragments different in C and B' is defined as CH=ch respectively
1ch
2ch
mand BR=br
1br
2br
n, concrete steps are as follows:
Step 5.1 supposes ch
ifor i-th participle in CH, br
jfor the participle of jth in BR, initial value i, j is set to 1
Step 5.2 compares ch respectively
iand br
jif,
illustrate that in first participle, braille participle comprises Chinese word segmentation, then for first participle, adopt the result br of braille participle
j; Contrary, if
then adopt the result ch of Chinese word segmentation
i
Step 5.3 initial setting up k=1
5.3.1 for
situation, definition ch
i, i+k=ch
ich
i+k, compare ch
i, i+kand br
j:
If a) ch
i, i+1=br
j, i=i+2 is set, j=j+1, if i>m or j>n, jumps to step 5.4, otherwise, jump to step 5.2
If b)
k=k+1, jumps to 5.3.1
If c)
ch is described
i+kin comprise br
jin the last character, the position defining this word is pos, then with pos be boundary, by ch
i+kbe divided into ch
i+k, posand ch
i+k, after_pos, wherein ch
i+k=ch
i+k, posch
i+k, after_pos, ch
i+k, posrepresent ch
i+kin the phrase of the 1st to pos word composition, ch
i+k, after_posrepresent ch
i+kin the phrase that forms to the last character of pos+1 word.By the i-th+k in Chinese word segmentation participle ch
i+k, after_posreplace, namely upgrade CH=ch
1ch
i+k-1ch
i+k, after_posch
i+k+1ch
m, i=i+k, j=j+1, jump to step 5.2
5.3.2 for
situation, definition br
j, j+k=br
jbr
j+k, compare br
j, j+kand ch
i:
If a) br
j, j+1=ch
i, then i=i+1, j=j+2, jumps to step 5.2
If b)
k=k+1, jumps to 5.3.2
If c)
br is described
j+kin comprise ch
iin the last character, the position defining this word is pos, then with pos be boundary, by br
j+kbe divided into br
j+k, posand br
j+k, after_pos, wherein br
j+k=br
j+k, posbr
j+k, after_pos, br
j+k, posrepresent br
j+kin the phrase of the 1st to pos word composition, br
j+k, after_posrepresent br
j+kin the phrase that forms to the last character of pos+1 word.By jth+k in braille participle participle br
j+k, after_posreplace, namely upgrade BR=br
1br
j+k-1br
j+k, after_posbr
j+k+1br
n, i=i+1, j=j+k, jump to step 5.2
Step 5.4 terminates integration algorithm
Step 6. is according to braille word link writing rule adjustment word segmentation result.Check the part of speech that participle is corresponding successively, and compare with the activation condition in braille word link writing rule set, if met, then the Conditions On The Results that applying rules is concentrated carries out participle or write the two or more syllables of a word together.Braille word link writing rule set form is as follows:
Combination handwriting rule: POS
k: [m, n]: POS
k-m+ ... + POS
k+ ... + POS
k+n→ POS
k-mpOS
k+n
Word segmentation regulation:
For the rule in rule set, the part of speech POS before first colon
kbe activation condition, can follow a bracket after rule, m and n of the inside represents front m the word and n word that need to check respectively current participle, if m and n is 0, then represents that this is a word segmentation regulation.Represent after second colon be participle part of speech combination, if meet this combination, then perform the operation after right arrow.
Step 7. braille mark is adjusted.Check the phonetic of the corresponding word of each participle successively, and the rule in assembling with braille mark is compared, if satisfied condition, then mark is carried out to current word and adjust.The form that braille mark is assembled is as follows:
Mark adjusts rule: tone
k: [n]: tone
ktone
k+n
Wherein tone
kfor the phonetic of current word, the n in square bracket represents the phonetic of rear n the word needing to check current word, tone
ktone
k+nfor mark tune condition, if pinyin sequence meets mark tune condition, then to tone
kcarry out mark to adjust
Step 8. braille shows, and outputs on blind aobvious device by braille.Current existing various point can be adopted to show device product, and call its corresponding output interface.
The present invention also proposes a kind of system reading Chinese character for blind person, comprising:
Obtain pinyin string module, for obtaining Chinese language text, participle operation is carried out to described Chinese language text, generate Chinese character string, by pronunciation dictionary, polyphone dictionary and word frequency information, with reference to the part-of-speech tagging that participle obtains, each word in described Chinese character string be converted to corresponding phonetic and connect for pinyin string;
Obtain new braille participle and adjusting module, for the contrast dictionary by searching phonetic and blind symbol, described pinyin string is converted to blind symbol string, by participle model, braille participle is carried out to described blind symbol string, generate initial braille participle, described Chinese character string and described initial braille participle are merged, generates new braille participle, according to braille word link writing rule, described new braille participle is adjusted;
Braille display module, for carrying out braille mark tune to according to the described new braille participle after braille word link writing rule adjustment, generates final braille participle, is shown by described final braille participle.
The concrete steps in described acquisition pinyin string module, described Chinese character string being converted to pinyin string are:
Step 2.1, for each word in described Chinese character string, judges whether each word is multi-character words, if multi-character words, and the phonetic that described multi-character words can be found in pronunciation dictionary corresponding, then directly return phonetic corresponding to described multi-character words, otherwise perform step 2.2;
Described multi-character words cutting is the sequence of Chinese character by step 2.2, gets Chinese characters all in described multi-character words successively, to each Chinese character, performs step 2.3 to 2.4;
Step 2.3, for current Chinese character, searches polyphone dictionary, judges whether described current Chinese character is polyphone, if not polyphone, searches the phonetic of described current Chinese character and return described phonetic in pronunciation dictionary; Otherwise perform step 2.4;
If step 2.4 polyphone, then perform following steps, concrete steps are:
If the current polyphone of step 2.4.1 comes from a monosyllabic word, then directly perform step 2.4.2; If multi-character words, then perform following step:
For the polyphone w in multi-character words
k, a) step, forms the word W of a n+1 word with a follow-up n word
k,n=w
kw
k+1w
k+n, in polyphone phrase dictionary, search W
k,n, as found, then with W
k,nin be searched the pronunciation of word as polyphone w
kpronunciation and return; As do not found, then performing b) step, forming the word W of a n+1 word with n word above
n-k, k=w
n-kw
n-kk+1w
n, in polyphone phrase dictionary, search W
n-k, k, as found, then with W
k,nin be searched word pronunciation as polyphone pronunciation and return, as do not searched, then respectively with follow-up and above n-1 word form the word W of a n word
k, n-1, W
n-k+1, k, described multi-character words to be performed a) respectively, b) step, until determine described polyphone w
kpronunciation;
Step 2.4.2 supposes that described polyphone has tone
1..., tone
nn pronunciation altogether, participle part of speech definition of probability is P
pos, weights are λ
1, probabilistic language model is defined as P
lm, weights are λ
2, participle word frequency definition of probability is P
freq, weights are λ
3, system is that each pronunciation of described polyphone calculates a score Score
i, wherein Score
i=λ
1p
pos(tone
i)+λ
2p
lm(tone
i)+λ
3p
freq(tone
i), take out the pronunciation that score is the highest
as polyphone final phonetic and return.
The new braille participle of described acquisition the step of carrying out in adjusting module merging are, for described Chinese character string C=c
1c
2c
mwith described initial braille participle B=b
1b
2b
n, wherein c
i, b
jrepresent a participle in described Chinese character string and described initial braille participle respectively, for described initial braille participle B, B is mapped to corresponding described Chinese character string B'=b
1' b'
2b'
n, wherein b'
jfor described initial braille participle b
jbe mapped as the participle after Chinese.
Described acquisition new braille participle and in adjusting module braille word link writing rule as follows:
Combination handwriting rule: POS
k: [m, n]: POS
k-m+ ... + POS
k+ ... + POS
k+n→ POS
k-mpOS
k+n
Word segmentation regulation:
POS
kfor activation condition, m and n represents front m the word and n word that need to check current new braille participle respectively, if m and n is 0, then represent that this is a word segmentation regulation, represent after second colon be participle part of speech combination, if meet this combination, then perform the operation after right arrow.
The concrete steps that braille mark described in described braille display module is adjusted are:
Check the phonetic of the corresponding word of the described new braille participle after each adjustment successively, and the rule in assembling with braille mark is compared, if satisfied condition, then carry out mark to current new braille participle and adjust, the form that described braille mark is assembled is as follows:
Mark adjusts rule: tone
k: [n]: tone
ktone
k+n
Wherein tone
kfor the phonetic of current new braille participle, n is the phonetic needing to check rear n of current new braille participle new braille participle, tone
ktone
k+nfor mark tune condition, if pinyin sequence meets mark tune condition, then to tone
kcarry out mark to adjust.
Below by a Chinese sentence is carried out Chinese to the conversion of braille and display as an example, introduce the implementation process reading the method and system of Chinese character for blind person of the present invention in detail, should be understood that this example is just for illustrating, instead of intention limits the scope of the invention.
Suppose that the Chinese sentence that need be converted to braille is: " Beijing be their destination ", adopt Chinese word segmenting module carry out Chinese word segmenting and carry out part-of-speech tagging, the result obtained is: " Beijing/NR is /VC they/PN/DEG object/NN ground/NN ".
Calling Chinese character string, to pinyin string modular converter, word segmentation result is converted to pinyin string, for " Beijing ", "Yes", " they ", " object " these five words, directly can confirm pronunciation by searching Pronounceable dictionary; For " " and " " these two words, owing to being all polyphone, the pronunciation of algorithm determination polyphone need be called.
For " " word, by part-of-speech tagging known " " part of speech of word is " DEG ", can confirm that the pronunciation of this word is for " de " by " DEG ", owing to can uniquely be confirmed by part of speech " " word pronunciation, so:
P
pos(de)＝1,
P
pos(di)＝0
Be under the condition of " they " at previous word, by searching probabilistic language model, the probability that can obtain for " de " is 0.45, pronounces for the probability of " di " is 0.05:
P
lm(de)＝P(de|tamen)＝0.45
P
lm(di)＝P(di|tamen)＝0.05
After being normalized, can obtain: P
lm(de)=0.9, P
lm(di)=0.1
Search in word frequency dictionary " " individual character word frequency, pronounce for the number of times of " de " is 185 times, pronounce for the number of times of " di " is 75 times, known by calculating, pronounce for the probability of " de " is 0.71, pronounce for the probability of " di " is 0.29
Based on experience value, arrange part of speech, language model, word frequency three probability weight be all 1/3, then:
Compared by score, can determine polyphone " " final pronunciation for " de ".
Similar, can determine " " pronunciation of word is " di ".Finally obtain pinyin string corresponding to Chinese sentence for " beijingshitamendemudidi ".
Call pinyin string to blind symbol string modular converter, obtain blind symbol string corresponding to pinyin string be " B! G*:T9M0DMUDIDI ".(braille adopted in this instructions is expressed as the ASCII character coding of blind symbol, and the some position form of non-blind symbol.Hereinafter identical.)
Call braille word-dividing mode and participle carried out to blind symbol string, obtain the string of the blind symbol after participle for " B! G*:|T9M0|D|MUDIDI ".
Call Chinese and braille word segmentation result Fusion Module merges Chinese word segmentation result and braille word segmentation result.Braille string after participle is corresponded to Chinese string, can obtain adopting the Chinese character string of braille participle for " Beijing is/they// destination ", the Chinese character string of braille participle and the Chinese character string of Chinese word segmenting be carried out editing distance and aligns, subordinate list 1 can be obtained:
Subordinate list 1: Chinese, the braille participle table of comparisons
Chinese and braille participle in contrast subordinate list 1, have the fragment that two different, fragment 1 " Beijing is " and fragment 2 " destination ".
Fragment 1 is processed, the Chinese word segmenting of fragment 1 is " Beijing/be ", braille participle is " Beijing is ", first participle " Beijing is " of getting Chinese word segmenting first participle " Beijing " and braille participle contrasts, because in braille participle, first word " Beijing is " contains first word " Beijing " in Chinese word segmenting, continue second word "Yes" of checking Chinese word segmenting, and carry out being combined to form " Beijing is " with first word " Beijing " and contrast with first word " Beijing is " of braille participle, because both are identical and no longer include other untreated word in fragment 1, according to choosing the rule of the more word of number of words as final participle, therefore determine that the participle of fragment 1 is for " Beijing is ".
Similar, can determine that the participle of fragment 2 is for " destination ".Finally, can determine that the word segmentation result after merging is for " Beijing is/they// destination ".
Call word segmentation result adjusting module, according to Chinese word segmenting annotation results, Pekinese's part of speech is " NR ", i.e. proper noun, for proper noun in braille standard, heel single syllable generic noun just carries out write the two or more syllables of a word together, " Beijing " heel "Yes" in example, part of speech is " VC ", i.e. " link-verb ", do not meet the condition of braille standard, write the two or more syllables of a word together should not be carried out, the participle " Beijing is " that reply is merged splits, obtain in " Beijing/be ", after adjustment, the word segmentation result obtained for " Beijing/be/they// destination ", the braille participle representation of its correspondence be " B! G*:T9M0DMUDIDI ".
Call braille mark mode transfer block and mark tune is carried out to word segmentation result.Specify in braille standard, " he ", " she ", " word " need use special method for expressing, must mark tune for " she " word.The blind symbol of " she " is " T9 ", and tone is first sound, being expressed as " A " in blind symbol, mark adjust after the representation of braille string be " B! G*:T9AM0DMUDIDI ".
Call braille display module braille string is presented on blind aobvious device.
Claims (10)
1. read a method for Chinese character for blind person, it is characterized in that, comprising:
Step 1, obtains Chinese language text, carries out participle operation to described Chinese language text, generate Chinese character string, by pronunciation dictionary, polyphone dictionary and word frequency information, with reference to the part-of-speech tagging that participle obtains, each word in described Chinese character string be converted to corresponding phonetic and connect for pinyin string;
Step 2, by searching the contrast dictionary of phonetic and blind symbol, described pinyin string is converted to blind symbol string, by participle model, braille participle is carried out to described blind symbol string, generate initial braille participle, described Chinese character string and described initial braille participle are merged, generates new braille participle, according to braille word link writing rule, described new braille participle is adjusted;
Step 3, carries out braille mark tune to according to the described new braille participle after braille word link writing rule adjustment, generates final braille participle, is shown by described final braille participle.
2. the method reading Chinese character for blind person as claimed in claim 1, it is characterized in that, the concrete steps in described step 1, described Chinese character string being converted to pinyin string are:
Step 2.1, for each word in described Chinese character string, judges whether each word is multi-character words, if multi-character words, and the phonetic that described multi-character words can be found in pronunciation dictionary corresponding, then directly return phonetic corresponding to described multi-character words, otherwise perform step 2.2;
Described multi-character words cutting is the sequence of Chinese character by step 2.2, gets Chinese characters all in described multi-character words successively, to each Chinese character, performs step 2.3 to 2.4;
Step 2.3, for current Chinese character, searches polyphone dictionary, judges whether described current Chinese character is polyphone, if not polyphone, searches the phonetic of described current Chinese character and return described phonetic in pronunciation dictionary; Otherwise perform step 2.4;
If step 2.4 polyphone, then perform following steps, concrete steps are:
If the current polyphone of step 2.4.1 comes from a monosyllabic word, then directly perform step 2.4.2; If multi-character words, then perform following step:
For the polyphone w in multi-character words
k, a) step, forms the word W of a n+1 word with a follow-up n word
k,n=w
kw
k+1w
k+n, in polyphone phrase dictionary, search W
k,n, as found, then with W
k,nin be searched the pronunciation of word as polyphone w
kpronunciation and return; As do not found, then performing b) step, forming the word W of a n+1 word with n word above
n-k, k=w
n-kw
n-kk+1w
n, in polyphone phrase dictionary, search W
n-k, k, as found, then with W
k,nin be searched word pronunciation as polyphone pronunciation and return, as do not searched, then respectively with follow-up and above n-1 word form the word W of a n word
k, n-1, W
n-k+1, k, described multi-character words to be performed a) respectively, b) step, until determine described polyphone w
kpronunciation;
Step 2.4.2 supposes that described polyphone has tone
1..., tone
nn pronunciation altogether, participle part of speech definition of probability is P
pos, weights are λ
1, probabilistic language model is defined as P
lm, weights are λ
2, participle word frequency definition of probability is P
freq, weights are λ
3, system is that each pronunciation of described polyphone calculates a score Score
i, wherein Score
i=λ
1p
pos(tone
i)+λ
2p
lm(tone
i)+λ
3p
freq(tone
i), take out the pronunciation that score is the highest
as polyphone final phonetic and return.
3. the method reading Chinese character for blind person as claimed in claim 1, it is characterized in that, the step carrying out in described step 2 merging is, for described Chinese character string C=c
1c
2c
mwith described initial braille participle B=b
1b
2b
n, wherein c
i, b
jrepresent a participle in described Chinese character string and described initial braille participle respectively, for described initial braille participle B, B is mapped to corresponding described Chinese character string B '=b '
1b '
2b '
n, wherein b '
jfor described initial braille participle b
jbe mapped as the participle after Chinese.
4. the method reading Chinese character for blind person as claimed in claim 1, is characterized in that, in described step 2, braille word link writing rule is as follows:
Combination handwriting rule: POS
k: [m, n]: POS
k-m+ ... + POS
k+ ... + POS
k+n→ POS
k-mpOS
k+n
Word segmentation regulation:
POS
kfor activation condition, m and n represents front m the word and n word that need to check current new braille participle respectively, if m and n is 0, then represent that this is a word segmentation regulation, represent after second colon be participle part of speech combination, if meet this combination, then perform the operation after right arrow.
5. the method reading Chinese character for blind person as claimed in claim 1, is characterized in that, the concrete steps that braille mark described in described step 3 is adjusted are:
Check the phonetic of the corresponding word of the described new braille participle after each adjustment successively, and the rule in assembling with braille mark is compared, if satisfied condition, then carry out mark to current new braille participle and adjust, the form that described braille mark is assembled is as follows:
Mark adjusts rule: tone
k: [n]: tone
ktone
k+n
Wherein tone
kfor the phonetic of current new braille participle, n is the phonetic needing to check rear n of current new braille participle new braille participle, tone
ktone
k+nfor mark tune condition, if pinyin sequence meets mark tune condition, then to tone
kcarry out mark to adjust.
6. read a system for Chinese character for blind person, it is characterized in that, comprising:
Obtain pinyin string module, for obtaining Chinese language text, participle operation is carried out to described Chinese language text, generate Chinese character string, by pronunciation dictionary, polyphone dictionary and word frequency information, with reference to the part-of-speech tagging that participle obtains, each word in described Chinese character string be converted to corresponding phonetic and connect for pinyin string;
Obtain new braille participle and adjusting module, for the contrast dictionary by searching phonetic and blind symbol, described pinyin string is converted to blind symbol string, by participle model, braille participle is carried out to described blind symbol string, generate initial braille participle, described Chinese character string and described initial braille participle are merged, generates new braille participle, according to braille word link writing rule, described new braille participle is adjusted;
Braille display module, for carrying out braille mark tune to according to the described new braille participle after braille word link writing rule adjustment, generates final braille participle, is shown by described final braille participle.
7. the system reading Chinese character for blind person as claimed in claim 6, it is characterized in that, the concrete steps in described acquisition pinyin string module, described Chinese character string being converted to pinyin string are:
Step 2.1, for each word in described Chinese character string, judges whether each word is multi-character words, if multi-character words, and the phonetic that described multi-character words can be found in pronunciation dictionary corresponding, then directly return phonetic corresponding to described multi-character words, otherwise perform step 2.2;
Described multi-character words cutting is the sequence of Chinese character by step 2.2, gets Chinese characters all in described multi-character words successively, to each Chinese character, performs step 2.3 to 2.4;
Step 2.3, for current Chinese character, searches polyphone dictionary, judges whether described current Chinese character is polyphone, if not polyphone, searches the phonetic of described current Chinese character and return described phonetic in pronunciation dictionary; Otherwise perform step 2.4;
If step 2.4 polyphone, then perform following steps, concrete steps are:
If the current polyphone of step 2.4.1 comes from a monosyllabic word, then directly perform step 2.4.2; If multi-character words, then perform following step:
For the polyphone w in multi-character words
k, a) step, forms the word W of a n+1 word with a follow-up n word
k,n=w
kw
k+1w
k+n, in polyphone phrase dictionary, search W
k,n, as found, then with W
k,nin be searched the pronunciation of word as polyphone w
kpronunciation and return; As do not found, then performing b) step, forming the word W of a n+1 word with n word above
n-k, k=w
n-kw
n-kk+1w
n, in polyphone phrase dictionary, search W
n-k, k, as found, then with W
k,nin be searched word pronunciation as polyphone pronunciation and return, as do not searched, then respectively with follow-up and above n-1 word form the word W of a n word
k, n-1, W
n-k+1, k, described multi-character words to be performed a) respectively, b) step, until determine described polyphone w
kpronunciation;
Step 2.4.2 supposes that described polyphone has tone
1..., tone
nn pronunciation altogether, participle part of speech definition of probability is P
pos, weights are λ
1, probabilistic language model is defined as P
lm, weights are λ
2, participle word frequency definition of probability is P
freq, weights are λ
3, system is that each pronunciation of described polyphone calculates a score Score
i, wherein Score
i=λ
1p
pos(tone
i)+λ
2p
lm(tone
i)+λ
3p
freq(tone
i), take out the pronunciation that score is the highest
as polyphone final phonetic and return.
8. the as claimed in claim 6 system reading Chinese character for blind person, is characterized in that, the new braille participle of described acquisition the step of carrying out in adjusting module merging are, for described Chinese character string C=c
1c
2c
mwith described initial braille participle B=b
1b
2b
n, wherein c
i, b
jrepresent a participle in described Chinese character string and described initial braille participle respectively, for described initial braille participle B, B is mapped to corresponding described Chinese character string B '=b '
1b '
2b '
n, wherein b '
jfor described initial braille participle b
jbe mapped as the participle after Chinese.
9. the as claimed in claim 6 system reading Chinese character for blind person, is characterized in that, described acquisition new braille participle and in adjusting module braille word link writing rule as follows:
Combination handwriting rule: POS
k: [m, n]: POS
k-m+ ... + POS
k+ ... + POS
k+n→ POS
k-mpOS
k+n
Word segmentation regulation:
POS
kfor activation condition, m and n represents front m the word and n word that need to check current new braille participle respectively, if m and n is 0, then represent that this is a word segmentation regulation, represent after second colon be participle part of speech combination, if meet this combination, then perform the operation after right arrow.
10. the system reading Chinese character for blind person as claimed in claim 6, is characterized in that, the concrete steps that braille mark described in described braille display module is adjusted are:
Check the phonetic of the corresponding word of the described new braille participle after each adjustment successively, and the rule in assembling with braille mark is compared, if satisfied condition, then carry out mark to current new braille participle and adjust, the form that described braille mark is assembled is as follows:
Mark adjusts rule: tone
k: [n]: tone
ktone
k+n
Wherein tone
kfor the phonetic of current new braille participle, n is the phonetic needing to check rear n of current new braille participle new braille participle, tone
ktone
k+nfor mark tune condition, if pinyin sequence meets mark tune condition, then to tone
kcarry out mark to adjust.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510623525.5A CN105404621B (en) | 2015-09-25 | 2015-09-25 | A kind of method and system that Chinese character is read for blind person |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510623525.5A CN105404621B (en) | 2015-09-25 | 2015-09-25 | A kind of method and system that Chinese character is read for blind person |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105404621A true CN105404621A (en) | 2016-03-16 |
| CN105404621B CN105404621B (en) | 2018-07-10 |
Family
ID=55470115
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510623525.5A Active CN105404621B (en) | 2015-09-25 | 2015-09-25 | A kind of method and system that Chinese character is read for blind person |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105404621B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107203508A (en) * | 2016-03-17 | 2017-09-26 | 富士施乐实业发展（中国）有限公司 | Braille document generating method and system |
| CN107273357A (en) * | 2017-06-14 | 2017-10-20 | 北京百度网讯科技有限公司 | Modification method, device, equipment and the medium of participle model based on artificial intelligence |
| CN107368474A (en) * | 2017-07-07 | 2017-11-21 | 浙江理工大学 | A kind of automatical and efficient translation conversion method of Chinese to braille |
| CN107886808A (en) * | 2017-11-03 | 2018-04-06 | 中国科学院计算技术研究所 | A kind of braille point auxiliary mask method and system |
| CN108052936A (en) * | 2017-11-03 | 2018-05-18 | 中国科学院计算技术研究所 | A kind of braille image wing drop bearing calibration and system |
| CN108062886A (en) * | 2017-11-03 | 2018-05-22 | 中国科学院计算技术研究所 | Braille point interactive mode mask method and system |
| CN108461111A (en) * | 2018-03-16 | 2018-08-28 | 重庆医科大学 | Chinese medical treatment text duplicate checking method and device, electronic equipment, computer read/write memory medium |
| CN108491441A (en) * | 2018-02-12 | 2018-09-04 | 北京联合大学 | A kind of braille information statistic system |
| CN110920268A (en) * | 2019-11-19 | 2020-03-27 | 西安交通大学 | Braille inscription method and system |
| CN111078898A (en) * | 2019-12-27 | 2020-04-28 | 出门问问信息科技有限公司 | Polyphone marking method and device and computer readable storage medium |
| CN113035026A (en) * | 2021-03-10 | 2021-06-25 | 之江实验室 | Audio-visual tactile perception matching method without barriers for braille information |
| CN116432603A (en) * | 2023-03-27 | 2023-07-14 | 之江实验室 | Memory and calculation integrated Chinese braille chip |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1323003A (en) * | 2001-06-22 | 2001-11-21 | 清华大学 | Intelligent Chinese computer system for the blind |
| CN1323004A (en) * | 2001-06-08 | 2001-11-21 | 清华大学 | Automatic conversion method from Chinese braille to Chinese character |
| WO2002006916A2 (en) * | 2000-07-18 | 2002-01-24 | Yishay Langenthal | Reading aid for the blind |
| CN1591414A (en) * | 2004-06-03 | 2005-03-09 | 华建电子有限责任公司 | Automatic translating converting method for Chinese language to braille |
| CN102184172A (en) * | 2011-05-10 | 2011-09-14 | 中国科学院计算技术研究所 | Chinese character reading system and method for blind people |
-
2015
- 2015-09-25 CN CN201510623525.5A patent/CN105404621B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002006916A2 (en) * | 2000-07-18 | 2002-01-24 | Yishay Langenthal | Reading aid for the blind |
| CN1323004A (en) * | 2001-06-08 | 2001-11-21 | 清华大学 | Automatic conversion method from Chinese braille to Chinese character |
| CN1323003A (en) * | 2001-06-22 | 2001-11-21 | 清华大学 | Intelligent Chinese computer system for the blind |
| CN1591414A (en) * | 2004-06-03 | 2005-03-09 | 华建电子有限责任公司 | Automatic translating converting method for Chinese language to braille |
| CN102184172A (en) * | 2011-05-10 | 2011-09-14 | 中国科学院计算技术研究所 | Chinese character reading system and method for blind people |
Non-Patent Citations (4)
| Title |
|---|
| 朱小燕，包塔: "EasyBraille:中文汉语盲文自动转换系统", 《自然语言理解与机器翻译——全国第六届计算语言学联合学术会议论文集》 * |
| 李宏乔 等: "汉语—盲文机器翻译系统的研究与实现", 《计算机应用》 * |
| 杨潮，车磊: "汉字—盲文转换系统的设计", 《北京印刷学院学报》 * |
| 苏晨: "面向统计机器翻译的领域自适应方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107203508A (en) * | 2016-03-17 | 2017-09-26 | 富士施乐实业发展（中国）有限公司 | Braille document generating method and system |
| US10664659B2 (en) | 2017-06-14 | 2020-05-26 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method for modifying segmentation model based on artificial intelligence, device and storage medium |
| CN107273357A (en) * | 2017-06-14 | 2017-10-20 | 北京百度网讯科技有限公司 | Modification method, device, equipment and the medium of participle model based on artificial intelligence |
| CN107273357B (en) * | 2017-06-14 | 2020-11-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based word segmentation model correction method, device, equipment and medium |
| CN107368474A (en) * | 2017-07-07 | 2017-11-21 | 浙江理工大学 | A kind of automatical and efficient translation conversion method of Chinese to braille |
| CN107368474B (en) * | 2017-07-07 | 2020-08-04 | 浙江理工大学 | Automatic efficient translation and conversion method from Chinese to braille |
| CN108062886A (en) * | 2017-11-03 | 2018-05-22 | 中国科学院计算技术研究所 | Braille point interactive mode mask method and system |
| CN108052936A (en) * | 2017-11-03 | 2018-05-18 | 中国科学院计算技术研究所 | A kind of braille image wing drop bearing calibration and system |
| CN107886808B (en) * | 2017-11-03 | 2021-03-09 | 中国科学院计算技术研究所 | Braille square auxiliary labeling method and system |
| CN107886808A (en) * | 2017-11-03 | 2018-04-06 | 中国科学院计算技术研究所 | A kind of braille point auxiliary mask method and system |
| CN108491441B (en) * | 2018-02-12 | 2022-02-01 | 北京联合大学 | Braille information statistical system |
| CN108491441A (en) * | 2018-02-12 | 2018-09-04 | 北京联合大学 | A kind of braille information statistic system |
| CN108461111A (en) * | 2018-03-16 | 2018-08-28 | 重庆医科大学 | Chinese medical treatment text duplicate checking method and device, electronic equipment, computer read/write memory medium |
| CN110920268A (en) * | 2019-11-19 | 2020-03-27 | 西安交通大学 | Braille inscription method and system |
| CN110920268B (en) * | 2019-11-19 | 2021-05-28 | 西安交通大学 | Braille inscription method and system |
| CN111078898A (en) * | 2019-12-27 | 2020-04-28 | 出门问问信息科技有限公司 | Polyphone marking method and device and computer readable storage medium |
| CN111078898B (en) * | 2019-12-27 | 2023-08-08 | 出门问问创新科技有限公司 | Multi-tone word annotation method, device and computer readable storage medium |
| CN113035026A (en) * | 2021-03-10 | 2021-06-25 | 之江实验室 | Audio-visual tactile perception matching method without barriers for braille information |
| CN113035026B (en) * | 2021-03-10 | 2022-06-17 | 之江实验室 | Audio-visual tactile perception matching method without barriers for braille information |
| CN116432603A (en) * | 2023-03-27 | 2023-07-14 | 之江实验室 | Memory and calculation integrated Chinese braille chip |
| CN116432603B (en) * | 2023-03-27 | 2023-10-13 | 之江实验室 | Memory and calculation integrated Chinese braille chip |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105404621B (en) | 2018-07-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105404621A (en) | Method and system for blind people to read Chinese character | |
| CN106598939B (en) | A kind of text error correction method and device, server, storage medium | |
| WO2020186778A1 (en) | Error word correction method and device, computer device, and storage medium | |
| CN110717031A (en) | Intelligent conference summary generation method and system | |
| CN106847288A (en) | The error correction method and device of speech recognition text | |
| CN109637537B (en) | Method for automatically acquiring annotated data to optimize user-defined awakening model | |
| CN105957518A (en) | Mongolian large vocabulary continuous speech recognition method | |
| CN104166462A (en) | Input method and system for characters | |
| CN101788978B (en) | Chinese and foreign spoken language automatic translation method combining Chinese pinyin and character | |
| CN109241540A (en) | A kind of blind automatic switching method of Chinese based on deep neural network and system | |
| CN111832278B (en) | Document fluency detection method and device, electronic equipment and medium | |
| Singh | A computational phonetic model for indian language scripts | |
| CN111489746A (en) | Power grid dispatching voice recognition language model construction method based on BERT | |
| CN102214238A (en) | Device and method for matching similarity of Chinese words | |
| CN109754791A (en) | Acoustic-controlled method and system | |
| KR102297480B1 (en) | System and method for structured-paraphrasing the unstructured query or request sentence | |
| Jiang et al. | Braille to print translations for Chinese | |
| Wang et al. | Chinese to Braille translation based on Braille word segmentation using statistical model | |
| CN111429886B (en) | Voice recognition method and system | |
| CN111090720B (en) | Hot word adding method and device | |
| Dinarelli | Spoken language understanding: from spoken utterances to semantic structures | |
| Yadav et al. | Different Models of Transliteration-A Comprehensive Review | |
| Dinarelli et al. | Concept segmentation and labeling for conversational speech | |
| Akinwonmi | Development of a prosodic read speech syllabic corpus of the Yoruba language | |
| KR101777141B1 (en) | Apparatus and method for inputting chinese and foreign languages based on hun min jeong eum using korean input keyboard |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |