CN102855480A - Method and device for recognizing characters in image - Google Patents

Method and device for recognizing characters in image Download PDF

Info

Publication number
CN102855480A
CN102855480A CN2012102793687A CN201210279368A CN102855480A CN 102855480 A CN102855480 A CN 102855480A CN 2012102793687 A CN2012102793687 A CN 2012102793687A CN 201210279368 A CN201210279368 A CN 201210279368A CN 102855480 A CN102855480 A CN 102855480A
Authority
CN
China
Prior art keywords
knowledge
information
literal
inquiry
character area
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.)
Pending
Application number
CN2012102793687A
Other languages
Chinese (zh)
Inventor
韩钧宇
丁二锐
吴中勤
文林福
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Baidu Netcom Science and Technology Co Ltd
Original Assignee
Beijing Baidu Netcom Science and Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Baidu Netcom Science and Technology Co Ltd filed Critical Beijing Baidu Netcom Science and Technology Co Ltd
Priority to CN2012102793687A priority Critical patent/CN102855480A/en
Publication of CN102855480A publication Critical patent/CN102855480A/en
Pending legal-status Critical Current

Links

Images

Abstract

The invention provides a method and a device for recognizing characters in an image. The method includes: S1, obtaining a character area in the image to be recognized; S2, recognizing characters in the character area; S3, inquiring a knowledge base to obtain knowledge information corresponding to a recognition result by using the recognition result; and S4, pushing a knowledge expansion word package containing the knowledge information when returning to the recognition result. According to the method and the device, a user can obtain corresponding knowledge information while obtaining character recognition results in the image, further obtaining of the knowledge information through a manual mode is not needed; and obviously, the method and the device are convenient and labor-saving.

Description

A kind of image character recognition method and device
[technical field]
The present invention relates to the Computer Applied Technology field, particularly a kind of method and apparatus of pictograph identification.
[background technology]
Along with developing rapidly of mobile Internet, the application of the image that the movement-based terminal camera collects is more and more extensive.Wherein the pictograph recognition technology is identified the literal in the image, is converted to text, thereby has alleviated the burden that the user inputs corresponding Word message, makes things convenient for the user to store, edit corresponding Word message.
In actual application, there is following situation, in the image of user by the portable terminal shooting, having much is the literal that the user is not familiar with or the user understands, such as some rarely used words, poem etc., in this case, the user is except wanting to extract the knowledge information of also thinking further to understand literal these word contents, the conventional images character recognition technology then can't address this problem, the user need to further inquire about by manual mode, for example inquires about in queries dictionary or the manual inputted search engine.
[summary of the invention]
In view of this, the invention provides a kind of image character recognition method and device, so that make things convenient for the user to obtain the knowledge information of pictograph.
Concrete technical scheme is as follows:
A kind of method of pictograph identification, the method comprises:
S1, obtain the character area in the image to be identified;
S2, described character area is carried out literal identification;
S3, utilize the recognition result search knowledge base to obtain knowledge information corresponding to recognition result;
S4, when returning described recognition result, push the knowledge comprise described knowledge information and expand the word bag.
According to one preferred embodiment of the present invention, described step S1 specifically comprises:
The image to be identified that the server mobile terminal receive sends extracts character area from described image to be identified; Perhaps,
The character area that the server mobile terminal receive extracts and sends from image to be identified.
According to one preferred embodiment of the present invention, described step S2 specifically comprises:
Character area is carried out binaryzation;
Character area after the binaryzation is divided into each block;
Extract the characteristic information of each block and mate with property data base, with the recognition result of matching result as each block;
Recognition result with each block makes up the recognition result that obtains described character area in order.
According to one preferred embodiment of the present invention, described step S3 specifically comprises with lower a kind of or combination in any:
Inquiry literal dictionary obtains the translation information of pronunciation, implication, usage or other languages of literal;
Inquiry cultural knowledge storehouse obtains corresponding allusion or the source information of literal;
Inquiry books inquiry storehouse obtains corresponding book information or the network resource information of literal;
Inquiry merchandise query storehouse obtains merchandise news corresponding to literal;
Query music inquiry storehouse obtains corresponding music information or the network resource information of literal;
Requester network application searches storehouse obtains corresponding network application information or the Web portal of literal.
According to one preferred embodiment of the present invention, all knowledge bases of inquiry in described step S3 generate knowledge with all knowledge informations that obtain and expand the word bag; Perhaps,
Also obtain the personalization option content of user selection in described step S1, knowledge base corresponding to the described personalization option content of inquiry in described step S3 generates knowledge with the knowledge information that obtains and expands the word bag; Perhaps,
All knowledge bases of inquiry in described step S3 are further determined weight corresponding to knowledge information to the knowledge information that inquires, and the knowledge information that weighted value is come top n generates knowledge expansion word bag, and N is default positive integer.
According to one preferred embodiment of the present invention, weight corresponding to described knowledge information determined in the following ways:
Determine the weight that this knowledge information is corresponding according to the total degree that knowledge information is queried to, the larger weighted value of total degree is larger; Perhaps,
The total degree that utilizes knowledge information to be queried to is determined the knowledge weight that this knowledge information is corresponding, utilize the total degree that all knowledge informations of classification are checked by the active user under this knowledge information to determine the user individual weight, utilize the knowledge weight of knowledge information to determine the weight that this knowledge information is corresponding with the product of user individual weight.
A kind of device of pictograph identification, this device comprises:
The zone acquiring unit is for the character area that obtains image to be identified;
Word recognition unit is used for described character area is carried out literal identification;
The knowledge query unit, the recognition result search knowledge base that is used for described word recognition unit obtains knowledge information corresponding to recognition result;
Push unit is used for when returning described recognition result as a result, pushes the knowledge that comprises described knowledge information and expands the word bag.
According to one preferred embodiment of the present invention, the image to be identified that described regional acquiring unit mobile terminal receive sends extracts character area from described image to be identified; Perhaps, the mobile terminal receive character area that from image to be identified, extracts and send.
According to one preferred embodiment of the present invention, described word recognition unit is specifically carried out: character area is carried out binaryzation, character area after the binaryzation is divided into each block, extract the characteristic information of each block and mate with property data base, with the recognition result of matching result as each block, the recognition result with each block makes up the recognition result that obtains described character area in order.
According to one preferred embodiment of the present invention, the concrete execution with lower a kind of or combination in any in described knowledge query unit:
Inquiry literal dictionary obtains the translation information of pronunciation, implication, usage or other languages of literal;
Inquiry cultural knowledge storehouse obtains corresponding allusion or the source information of literal;
Inquiry books inquiry storehouse obtains corresponding book information or the network resource information of literal;
Inquiry merchandise query storehouse obtains merchandise news corresponding to literal;
Query music inquiry storehouse obtains corresponding music information or the network resource information of literal;
Requester network application searches storehouse obtains corresponding network application information or the Web portal of literal.
According to one preferred embodiment of the present invention, all knowledge bases are inquired about in described knowledge query unit, all knowledge informations that obtain are generated knowledge expand the word bag; Perhaps,
Described regional acquiring unit also is used for obtaining the personalization option content of user selection, and knowledge base corresponding to described personalization option content inquired about in described knowledge query unit, the knowledge information that obtains is generated knowledge expand the word bag; Perhaps,
All knowledge bases are inquired about in described knowledge query unit, and the knowledge information that inquires is further determined weight corresponding to knowledge information, and the knowledge information that weighted value is come top n generates knowledge expansion word bag, and N is default positive integer.
According to one preferred embodiment of the present invention, described knowledge query unit is determined weight corresponding to described knowledge information in the following ways:
Determine the weight that this knowledge information is corresponding according to the total degree that knowledge information is queried to, the larger weighted value of total degree is larger; Perhaps,
The total degree that utilizes knowledge information to be queried to is determined the knowledge weight that this knowledge information is corresponding, utilize the total degree that all knowledge informations of classification are checked by the active user under this knowledge information to determine the user individual weight, utilize the knowledge weight of knowledge information to determine the weight that this knowledge information is corresponding with the product of user individual weight.
As can be seen from the above technical solutions, the present invention utilizes the result queries knowledge base of literal identification to obtain knowledge information corresponding to recognition result, and be included in knowledge and expand in the word bag and return to the user together with recognition result, so that the user is when obtaining image Chinese word recognition result, can get access to corresponding knowledge information, and need not further to pass through manual mode acquire knowledge information, obviously convenient and laborsaving.
[description of drawings]
The method flow diagram that Fig. 1 identifies for the pictograph that the embodiment of the invention one provides;
The system schematic that Fig. 2 provides for the embodiment of the invention;
The structure drawing of device that Fig. 3 identifies for the pictograph that the embodiment of the invention two provides;
Two bandwagon effect schematic diagram of the portable terminal that Fig. 4 and Fig. 5 provide for the embodiment of the invention.
[embodiment]
In order to make the purpose, technical solutions and advantages of the present invention clearer, describe the present invention below in conjunction with the drawings and specific embodiments.
Embodiment one,
The method flow diagram that Fig. 1 identifies for the pictograph that the embodiment of the invention provides, as shown in Figure 1, the method can may further comprise the steps:
Step 101: obtain the character area in the image to be identified.
Server obtains the image that comprises Word message that portable terminal sends, and this image can be the original image that portable terminal photographs, and server extracts the character area in the image to be identified in this step.Perhaps, this image can be after portable terminal photographs original image, the character area in the image to be identified is sent to server after extracting the character area in the image to be identified.
When extracting character area, can adopt existing mode, extract character area after removing image background, can adopt but be not limited to following mode:
Mode one, at first carry out color run-length coding according to colored Euclidean distance, then carry out color cluster, carry out generation and the selection of character layer based on cluster result, for example Retention area is greater than the connected domain of certain value, Euclidean distance based on connected domain and each color cluster center generates each image aspect, at last determine literal aspect, noise aspect or background aspect according to the relation of the number of pixels of the number of pixels of each image aspect and this layer segmentation threshold, just obtain the character layer face after taking out at last noise aspect and background aspect, i.e. character area.
Mode two, select a large amount of literal sample images and do not contain the picture of literal, use the marginal information of this two classes picture of canny operator extraction as the training sample of rarefaction representation classifying dictionary; Two class training samples input classification rarefaction representation dictionary training algorithm is obtained literal rarefaction representation classifying dictionary and non-legible rarefaction representation classifying dictionary; Transfer image to be identified to gray level image, use the marginal information of canny operator extraction gray level image; Utilization is based on the candidate character region in the rarefaction representation extraction gray-scale Image Edge information of classifying dictionary; Use respectively the distance of swimming smoothing algorithm edge that candidate character region is isolated to be connected to larger zone in the horizontal direction with on the vertical direction, carry out again Projection Analysis and find out corresponding literal line, cast out simultaneously the capable isolated edge in addition of candidate character region Chinese word; Detected character area is identified out.
If portable terminal carries out the extraction of character area, then can adopt existing character area extraction software or manual mode to carry out the extraction of character area.
In addition, the character area that obtains in this step can be one, also can be more than two.Because the content in this step is prior art, does not repeat them here.
Step 102: character area is carried out literal identification.
The process of wherein character area being carried out literal identification can adopt prior art equally, namely may further comprise the steps: character area is carried out binaryzation; Character area after the binaryzation is divided into each block; Extract the characteristic information of each block and mate with property data base, with the recognition result of matching result as each block, the recognition result with each block makes up the recognition result that just obtains character area in order.
In addition, the mode of literal identification is varied, can also adopt other can realize arbitrarily the mode of literal identification except aforesaid way, specifically repeats no more.
Step 103: utilize the recognition result search knowledge base to obtain knowledge information corresponding to recognition result.
The knowledge base of inquiring about in this step can include but not limited to a kind of or combination in any in the following knowledge base:
1) literal dictionary, inquiry this article word dictionary can obtain the translation information etc. of pronunciation, implication, usage or other languages of literal.For example the user photographs a rarely used word in the book by portable terminal, and server by utilizing recognition result inquiry literal dictionary just can obtain the information such as the pronunciation, implication, example sentence of this uncommon word.
2) allusion that this culture knowledge base can obtain literal, source information etc. are inquired about in cultural knowledge storehouse.For example, the user photographs one section poem by portable terminal, and server by utilizing recognition result inquiry cultural knowledge storehouse just can obtain the source of this poem.
3) books inquiry storehouse is inquired about this books inquiry storehouse and can be obtained book information corresponding to literal or network resource information etc.For example, the user photographs a book cover by portable terminal, the bookstore etc. that server by utilizing recognition result inquiry books inquiry storehouse just can obtain publishing house's information, the price of these books and sell these books.
4) merchandise query storehouse is inquired about this merchandise query storehouse and can be obtained merchandise news corresponding to literal, such as trade name, merchandise resources, commodity price, commodity points for attention, commodity evaluation etc.For example the user photographs a commodity sign by portable terminal, and server by utilizing recognition result inquiry merchandise query storehouse just can obtain to sell the bookstore of these commodity, the price of commodity etc.Again for example, the user takes a medicine, and server by utilizing recognition result inquiry commodity storehouse just can obtain the information such as dosage, especially in the points for attention of particular time.
5) music inquiry storehouse is inquired about this music inquiry storehouse and can be obtained music information corresponding to literal or site resource information etc.For example the user takes an album cover by portable terminal, server by utilizing recognition result inquiry merchandise query storehouse just can obtain this special edition song information, can audiovisual or download the site resource etc. of this patent.
6) network application search library is inquired about this network application search library and can be obtained network application information corresponding to literal or Web portal etc.For example the user photographs the picture of an app by portable terminal, and server by utilizing recognition result requester network application searches storehouse just can obtain the relevant information of this app and the Web portal of this app etc.
Above-mentioned knowledge base can be local knowledge base, also can be online networked knowledge base, also can be the knowledge base that provides by the open data-interface access third party of third party.
A kind of implementation is that after having inquired about all knowledge bases, all knowledge informations that coupling is obtained generate knowledge expansion word bag for returning to portable terminal.
Owing to may there be the knowledge base of a greater number, the user may not need the knowledge information of so much kind, therefore another kind of implementation is, portable terminal provides the personal settings option to the user, in the option content that when server sends image, sends simultaneously user selection, server is in this step during search knowledge base, only knowledge base corresponding to option content selected of inquiring user.
Give an example, portable terminal provides the literal dictionary to the user, cultural knowledge, the books inquiry, merchandise query, the music inquiry, the personal settings options such as network application, if the user has taken the image of uncommon word by portable terminal, can select this option of literal dictionary, then portable terminal sends to server with the option content of image and user selection, server is for the literal recognition result search knowledge base of image the time, just can only inquire about this knowledge base of literal dictionary, the Query Result that obtains is generated knowledge expand the word bag, then in step 104, return to portable terminal together with the literal recognition result.Certainly, the user can select more than one option.
Also have a kind of implementation, when search knowledge base, still inquire about all knowledge bases, but selectively return when the knowledge information of returning, the knowledge information of wherein selecting to return any or several kinds can be by the mode based on the weight ordering.Particularly, if in the knowledge base of certain classification, inquire knowledge information corresponding to recognition result, then further determine weight corresponding to this knowledge information, the generation knowledge expansion word bag that weighted value comes top n in the knowledge information that obtains the most at last is for returning to portable terminal, and N is default positive integer.
Wherein weight corresponding to knowledge information can adopt but be not limited to following mode and determine: the total degree that one, the knowledge information that this recognition result is corresponding are queried to, the larger weighted value of this total degree is larger.They are two years old, the total degree that utilizes knowledge information corresponding to recognition result to be inquired by all users is determined the knowledge weight that this knowledge information is corresponding, all knowledge informations of classification are checked by the active user (after soon this knowledge information is pushed to portable terminal under the recycling knowledge information, the user can check the wherein knowledge information of some classification, for example pushed corresponding book information and the merchandise news of certain Word message to the user, if the user has checked merchandise news wherein, then can upgrade the number of times that merchandise classification knowledge information is checked, be used for to upgrade the user individual weight of merchandise classification knowledge information) total degree determine the user individual weight, with the product of the knowledge weight of knowledge information and user individual weight as weight corresponding to this knowledge information.
In addition, this step can be based on whole Word messages of recognition result when search knowledge base, also can be based on recognition result being cut the crucial meaning Word message that obtains behind the word.
Step 104: when returning recognition result, push the knowledge that comprises corresponding knowledge information and expand the word bag.
After server was expanded recognition result and knowledge the word bag and returned to portable terminal, the user just can get access to the knowledge information of correspondence when recognition result is obtained in the demonstration of portable terminal.And, knowledge information wherein may be more than one classifications, if the user has checked wherein some or several classifications, then can report to server, by the total degree that each knowledge information of server update is inquired about by all users, user individual weight corresponding to Knowledge category under the information of refreshing one's knowledge simultaneously.
More than be the description that method provided by the present invention is carried out, be described in detail below by two pairs of devices provided by the present invention of embodiment.For convenient understanding at first is described the applied system of said method of the present invention, as shown in Figure 2, this system is made of portable terminal and server, wherein portable terminal can send to server as image to be identified with the image that comprises literal that photographs, therefrom extract character area by server, perhaps, the image that comprises literal that portable terminal will photograph as image to be identified after, therefrom extract character area, this character area is sent to server.Server is carried out flow process shown in the embodiment one afterwards.The device that the following embodiment two of the present invention provides is arranged in the server, is used for finishing flow process shown in the embodiment one.
Embodiment two,
The structure drawing of device of the pictograph identification that Fig. 3 provides for the embodiment of the invention two, as shown in Figure 3, this device comprises: regional acquiring unit 301, word recognition unit 302, knowledge query unit 303 and push unit 304 as a result.
At first, regional acquiring unit 301 obtains the character area in the image to be identified.
Herein, the image to be identified that regional acquiring unit 301 mobile terminal receives send extracts character area from image to be identified; Perhaps, the mobile terminal receive character area that from image to be identified, extracts and send.When extracting character area, can adopt the dual mode described in the step 101 among the embodiment one, because this partial content is prior art, be not described in detail in this.
Then 302 pairs of character areas of word recognition unit carry out literal identification.Concrete identifying can comprise: character area is carried out binaryzation, character area after the binaryzation is divided into each block, extract the characteristic information of each block and mate with property data base, with the recognition result of matching result as each block, the recognition result with each block makes up the recognition result that obtains character area in order.
Knowledge query unit 303 utilizes the recognition result search knowledge base of word recognition unit 302 to obtain knowledge information corresponding to recognition result.
Particularly, knowledge query unit 303 can be carried out when search knowledge base with lower a kind of or combination in any:
Inquiry literal dictionary obtains the translation information of pronunciation, implication, usage or other languages of literal;
Inquiry cultural knowledge storehouse obtains corresponding allusion or the source information of literal;
Inquiry books inquiry storehouse obtains corresponding book information or the network resource information of literal;
Inquiry merchandise query storehouse obtains merchandise news corresponding to literal;
Query music inquiry storehouse obtains corresponding music information or the network resource information of literal;
Requester network application searches storehouse obtains corresponding network application information or the Web portal of literal.
A kind of embodiment wherein, knowledge query unit 303 all knowledge bases of inquiry generate knowledge with all knowledge informations that obtain and expand word bags.
Another kind of embodiment, zone acquiring unit 301 also is used for obtaining the personalization option content of user selection, knowledge base corresponding to knowledge query unit 303 inquiry personalization option contents this moment generates knowledge with the knowledge information that obtains and expands word bag (not shown in this kind situation map 3).
Another embodiment, knowledge query unit 303 all knowledge bases of inquiry are further determined weight corresponding to knowledge information to the knowledge information that inquires, the knowledge information that weighted value is come top n generates knowledge expansion word bag.
Wherein can determine in the following ways the weight that knowledge information is corresponding:
Mode one, the total degree that is inquired by all users according to knowledge information are determined the weight that this knowledge information is corresponding, and the larger weighted value of total degree is larger.
Mode two, the total degree that utilizes knowledge information to be queried to are determined the knowledge weight that this knowledge information is corresponding, utilize the total degree that all knowledge informations of classification are checked by the active user under this knowledge information to determine the user individual weight, utilize the knowledge weight of knowledge information to determine the weight that this knowledge information is corresponding with the product of user individual weight.
At last, push unit 304 pushes the knowledge that comprises knowledge information and expands the word bag when returning recognition result as a result.
After recognition result and knowledge expanded the word bag and return to portable terminal, the user just can get access to the knowledge information of correspondence when recognition result is obtained in the demonstration of portable terminal.And, knowledge information wherein may be more than one classifications, if the user has checked wherein some or several classifications, then can report to server, upgrade the corresponding total degree that is inquired by all users by knowledge query unit 303 based on every knowledge information that the user checks, and user individual weight corresponding to the classification of refreshing one's knowledge.
By said method of the present invention and device, the user can get access to corresponding knowledge information when obtaining the pictograph recognition result, and need not further to pass through manual mode acquire knowledge information, and is obviously convenient and laborsaving.
For example, the user sees a rarely used word at book, by portable terminal it is sent to server after being filmed, carry out after the method for the present invention by server, when returning the literal recognition result to portable terminal, the knowledge informations such as pronunciation, implication and usage of this literal together can be sent to portable terminal.Its Chinese word recognition result and knowledge information are not limited at the exhibition method of portable terminal and invention, can adopt arbitrarily form, and near the form that for example frame is quoted in employing recognition result is showed knowledge information, as shown in Figure 4.
Again for example, comprise literal " The Book of Laughter and Forgetting " in the picture that the user takes, portable terminal is identified and knowledge query through literal after sending it to server, and server returns literal recognition result and knowledge information to portable terminal, can be as shown in Figure 5 in the displaying result of portable terminal.
The above only is preferred embodiment of the present invention, and is in order to limit the present invention, within the spirit and principles in the present invention not all, any modification of making, is equal to replacement, improvement etc., all should be included within the scope of protection of the invention.

Claims (12)

1. the method for a pictograph identification is characterized in that, the method comprises:
S1, obtain the character area in the image to be identified;
S2, described character area is carried out literal identification;
S3, utilize the recognition result search knowledge base to obtain knowledge information corresponding to recognition result;
S4, when returning described recognition result, push the knowledge comprise described knowledge information and expand the word bag.
2. method according to claim 1 is characterized in that, described step S1 specifically comprises:
The image to be identified that the server mobile terminal receive sends extracts character area from described image to be identified; Perhaps,
The character area that the server mobile terminal receive extracts and sends from image to be identified.
3. method according to claim 1 is characterized in that, described step S2 specifically comprises:
Character area is carried out binaryzation;
Character area after the binaryzation is divided into each block;
Extract the characteristic information of each block and mate with property data base, with the recognition result of matching result as each block;
Recognition result with each block makes up the recognition result that obtains described character area in order.
4. method according to claim 1 is characterized in that, described step S3 specifically comprises with lower a kind of or combination in any:
Inquiry literal dictionary obtains the translation information of pronunciation, implication, usage or other languages of literal;
Inquiry cultural knowledge storehouse obtains corresponding allusion or the source information of literal;
Inquiry books inquiry storehouse obtains corresponding book information or the network resource information of literal;
Inquiry merchandise query storehouse obtains merchandise news corresponding to literal;
Query music inquiry storehouse obtains corresponding music information or the network resource information of literal;
Requester network application searches storehouse obtains corresponding network application information or the Web portal of literal.
5. according to claim 1 or 4 described methods, it is characterized in that, all knowledge bases of inquiry in described step S3 generate knowledge with all knowledge informations that obtain and expand word bags; Perhaps,
Also obtain the personalization option content of user selection in described step S1, knowledge base corresponding to the described personalization option content of inquiry in described step S3 generates knowledge with the knowledge information that obtains and expands the word bag; Perhaps,
All knowledge bases of inquiry in described step S3 are further determined weight corresponding to knowledge information to the knowledge information that inquires, and the knowledge information that weighted value is come top n generates knowledge expansion word bag, and N is default positive integer.
6. method according to claim 5 is characterized in that, weight corresponding to described knowledge information determined in the following ways:
Determine the weight that this knowledge information is corresponding according to the total degree that knowledge information is queried to, the larger weighted value of total degree is larger; Perhaps,
The total degree that utilizes knowledge information to be queried to is determined the knowledge weight that this knowledge information is corresponding, utilize the total degree that all knowledge informations of classification are checked by the active user under this knowledge information to determine the user individual weight, utilize the knowledge weight of knowledge information to determine the weight that this knowledge information is corresponding with the product of user individual weight.
7. the device of a pictograph identification is characterized in that, this device comprises:
The zone acquiring unit is for the character area that obtains image to be identified;
Word recognition unit is used for described character area is carried out literal identification;
The knowledge query unit, the recognition result search knowledge base that is used for described word recognition unit obtains knowledge information corresponding to recognition result;
Push unit is used for when returning described recognition result as a result, pushes the knowledge that comprises described knowledge information and expands the word bag.
8. device according to claim 7 is characterized in that, the image to be identified that described regional acquiring unit mobile terminal receive sends extracts character area from described image to be identified; Perhaps, the mobile terminal receive character area that from image to be identified, extracts and send.
9. device according to claim 7, it is characterized in that, described word recognition unit is specifically carried out: character area is carried out binaryzation, character area after the binaryzation is divided into each block, extract the characteristic information of each block and mate with property data base, with the recognition result of matching result as each block, the recognition result with each block makes up the recognition result that obtains described character area in order.
10. device according to claim 7 is characterized in that, described knowledge query unit is concrete to be carried out with lower a kind of or combination in any:
Inquiry literal dictionary obtains the translation information of pronunciation, implication, usage or other languages of literal;
Inquiry cultural knowledge storehouse obtains corresponding allusion or the source information of literal;
Inquiry books inquiry storehouse obtains corresponding book information or the network resource information of literal;
Inquiry merchandise query storehouse obtains merchandise news corresponding to literal;
Query music inquiry storehouse obtains corresponding music information or the network resource information of literal;
Requester network application searches storehouse obtains corresponding network application information or the Web portal of literal.
11. according to claim 7 or 10 described devices, it is characterized in that, all knowledge bases are inquired about in described knowledge query unit, all knowledge informations that obtain are generated knowledge expand word bags; Perhaps,
Described regional acquiring unit also is used for obtaining the personalization option content of user selection, and knowledge base corresponding to described personalization option content inquired about in described knowledge query unit, the knowledge information that obtains is generated knowledge expand the word bag; Perhaps,
All knowledge bases are inquired about in described knowledge query unit, and the knowledge information that inquires is further determined weight corresponding to knowledge information, and the knowledge information that weighted value is come top n generates knowledge expansion word bag, and N is default positive integer.
12. device according to claim 11 is characterized in that, described knowledge query unit is determined weight corresponding to described knowledge information in the following ways:
Determine the weight that this knowledge information is corresponding according to the total degree that knowledge information is queried to, the larger weighted value of total degree is larger; Perhaps,
The total degree that utilizes knowledge information to be queried to is determined the knowledge weight that this knowledge information is corresponding, utilize the total degree that all knowledge informations of classification are checked by the active user under this knowledge information to determine the user individual weight, utilize the knowledge weight of knowledge information to determine the weight that this knowledge information is corresponding with the product of user individual weight.
CN2012102793687A 2012-08-07 2012-08-07 Method and device for recognizing characters in image Pending CN102855480A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2012102793687A CN102855480A (en) 2012-08-07 2012-08-07 Method and device for recognizing characters in image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2012102793687A CN102855480A (en) 2012-08-07 2012-08-07 Method and device for recognizing characters in image

Publications (1)

Publication Number Publication Date
CN102855480A true CN102855480A (en) 2013-01-02

Family

ID=47402057

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2012102793687A Pending CN102855480A (en) 2012-08-07 2012-08-07 Method and device for recognizing characters in image

Country Status (1)

Country Link
CN (1) CN102855480A (en)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103093217A (en) * 2013-01-06 2013-05-08 北京百度网讯科技有限公司 Interactive image and character recognition method and device
CN103294779A (en) * 2013-05-13 2013-09-11 北京百度网讯科技有限公司 Method and device for acquiring object information
CN103970753A (en) * 2013-01-28 2014-08-06 北大方正集团有限公司 Pushing method and pushing device for related knowledge
CN104134064A (en) * 2013-05-02 2014-11-05 百度国际科技(深圳)有限公司 Character recognition method and device
CN104143093A (en) * 2013-05-10 2014-11-12 百度在线网络技术(北京)有限公司 Character identification method and device
CN105426462A (en) * 2015-11-13 2016-03-23 深圳码隆科技有限公司 Image searching method and device based on image element
CN105447708A (en) * 2014-08-28 2016-03-30 阿里巴巴集团控股有限公司 Information offering method and information offering device
CN105631051A (en) * 2016-02-29 2016-06-01 华南理工大学 Character recognition based mobile augmented reality reading method and reading system thereof
CN105653733A (en) * 2016-02-26 2016-06-08 百度在线网络技术(北京)有限公司 Searching method and device
CN106021210A (en) * 2016-05-16 2016-10-12 珠海市魅族科技有限公司 Picture text processing method and apparatus
CN107291352A (en) * 2017-06-20 2017-10-24 广州阿里巴巴文学信息技术有限公司 Application program is redirected in a kind of word read method and its device
CN107438160A (en) * 2017-07-21 2017-12-05 上海明数数字出版科技有限公司 A kind of preview image scales the method and device into line character inquiry manually
CN111274505A (en) * 2018-12-05 2020-06-12 中版云教育科技(北京)有限公司 Resource viewing method and device
WO2021068628A1 (en) * 2019-10-11 2021-04-15 支付宝(杭州)信息技术有限公司 Image processing method and system
CN113221696A (en) * 2021-04-29 2021-08-06 四川大学华西医院 Image recognition method, system, equipment and storage medium
CN113359998A (en) * 2021-05-24 2021-09-07 维沃移动通信有限公司 Information query method and device
CN113704490A (en) * 2021-08-19 2021-11-26 扬州大学 New technology education knowledge graph construction method for heterogeneous data

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101000623A (en) * 2007-01-08 2007-07-18 深圳市宜搜科技发展有限公司 Method for image identification search by mobile phone photographing and device using the method
CN101044494A (en) * 2004-10-20 2007-09-26 摩托罗拉公司 An electronic device and method for visual text interpretation
US20080059526A1 (en) * 2006-09-01 2008-03-06 Sony Corporation Playback apparatus, searching method, and program
CN101751566A (en) * 2008-12-12 2010-06-23 汉王科技股份有限公司 Method and device for identifying and annotating menu based on handheld device

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101044494A (en) * 2004-10-20 2007-09-26 摩托罗拉公司 An electronic device and method for visual text interpretation
US20080059526A1 (en) * 2006-09-01 2008-03-06 Sony Corporation Playback apparatus, searching method, and program
CN101000623A (en) * 2007-01-08 2007-07-18 深圳市宜搜科技发展有限公司 Method for image identification search by mobile phone photographing and device using the method
CN101751566A (en) * 2008-12-12 2010-06-23 汉王科技股份有限公司 Method and device for identifying and annotating menu based on handheld device

Cited By (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103093217A (en) * 2013-01-06 2013-05-08 北京百度网讯科技有限公司 Interactive image and character recognition method and device
US9501587B2 (en) 2013-01-28 2016-11-22 Peking University Founder Group Co., Ltd. Method and device for pushing association knowledge
CN103970753A (en) * 2013-01-28 2014-08-06 北大方正集团有限公司 Pushing method and pushing device for related knowledge
CN103970753B (en) * 2013-01-28 2017-06-20 北大方正集团有限公司 The method for pushing and device of association knowledge
CN104134064A (en) * 2013-05-02 2014-11-05 百度国际科技(深圳)有限公司 Character recognition method and device
CN104143093A (en) * 2013-05-10 2014-11-12 百度在线网络技术(北京)有限公司 Character identification method and device
CN104143093B (en) * 2013-05-10 2018-01-09 百度在线网络技术(北京)有限公司 Character recognition method and device
CN103294779A (en) * 2013-05-13 2013-09-11 北京百度网讯科技有限公司 Method and device for acquiring object information
CN105447708A (en) * 2014-08-28 2016-03-30 阿里巴巴集团控股有限公司 Information offering method and information offering device
CN105426462A (en) * 2015-11-13 2016-03-23 深圳码隆科技有限公司 Image searching method and device based on image element
CN105653733A (en) * 2016-02-26 2016-06-08 百度在线网络技术(北京)有限公司 Searching method and device
CN105631051A (en) * 2016-02-29 2016-06-01 华南理工大学 Character recognition based mobile augmented reality reading method and reading system thereof
CN106021210A (en) * 2016-05-16 2016-10-12 珠海市魅族科技有限公司 Picture text processing method and apparatus
CN107291352A (en) * 2017-06-20 2017-10-24 广州阿里巴巴文学信息技术有限公司 Application program is redirected in a kind of word read method and its device
CN107438160A (en) * 2017-07-21 2017-12-05 上海明数数字出版科技有限公司 A kind of preview image scales the method and device into line character inquiry manually
CN111274505B (en) * 2018-12-05 2024-01-30 中版云教育科技(北京)有限公司 Resource viewing method and device
CN111274505A (en) * 2018-12-05 2020-06-12 中版云教育科技(北京)有限公司 Resource viewing method and device
WO2021068628A1 (en) * 2019-10-11 2021-04-15 支付宝(杭州)信息技术有限公司 Image processing method and system
CN113221696A (en) * 2021-04-29 2021-08-06 四川大学华西医院 Image recognition method, system, equipment and storage medium
CN113359998A (en) * 2021-05-24 2021-09-07 维沃移动通信有限公司 Information query method and device
CN113359998B (en) * 2021-05-24 2023-11-21 维沃移动通信有限公司 Information query method and device
CN113704490A (en) * 2021-08-19 2021-11-26 扬州大学 New technology education knowledge graph construction method for heterogeneous data
CN113704490B (en) * 2021-08-19 2023-06-20 扬州大学 New industrial and scientific education knowledge graph construction method for heterogeneous data

Similar Documents

Publication Publication Date Title
CN102855480A (en) Method and device for recognizing characters in image
US9042659B2 (en) Method and system for fast and robust identification of specific product images
US11899681B2 (en) Knowledge graph building method, electronic apparatus and non-transitory computer readable storage medium
CN102831200A (en) Commodity propelling method and device based on image character recognition
KR101806169B1 (en) Method, apparatus, system and computer program for offering a shopping information
WO2017066543A1 (en) Systems and methods for automatically analyzing images
CN106970991B (en) Similar application identification method and device, application search recommendation method and server
CN103988202A (en) Image attractiveness based indexing and searching
CN105631051A (en) Character recognition based mobile augmented reality reading method and reading system thereof
CN102833625A (en) Device and method for dynamically embedding advertisement into video
CN110110090A (en) Searching method, education search engine system and device
CN105893404A (en) Natural information identification based pushing system and method, and client
CN111522901A (en) Method and device for processing address information in text
CN107636646B (en) Facility anchoring with geo-located imaging
US20130191323A1 (en) System and method for identifying the context of multimedia content elements displayed in a web-page
US9906588B2 (en) Server and method for extracting content for commodity
CN105243120A (en) Retrieval method and apparatus
CN103412880A (en) Method and device for determining implicit associated information between multimedia resources
US20150205792A1 (en) Method for object recognition, corresponding system, apparatus and computer program product
CN110363206B (en) Clustering of data objects, data processing and data identification method
CN114416939A (en) Intelligent question and answer method, device, equipment and storage medium
JPWO2007004521A1 (en) Marker specifying device and marker specifying method
CN111881900B (en) Corpus generation method, corpus translation model training method, corpus translation model translation method, corpus translation device, corpus translation equipment and corpus translation medium
CN111638792A (en) AR effect presentation method and device, computer equipment and storage medium
CN103377199B (en) Information processor and information processing method

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
EXSB Decision made by sipo to initiate substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20130102

RJ01 Rejection of invention patent application after publication