CN102831200A - Commodity propelling method and device based on image character recognition - Google Patents
Commodity propelling method and device based on image character recognition Download PDFInfo
- Publication number
- CN102831200A CN102831200A CN2012102793672A CN201210279367A CN102831200A CN 102831200 A CN102831200 A CN 102831200A CN 2012102793672 A CN2012102793672 A CN 2012102793672A CN 201210279367 A CN201210279367 A CN 201210279367A CN 102831200 A CN102831200 A CN 102831200A
- Authority
- CN
- China
- Prior art keywords
- commodity
- commodity information
- character
- weight
- information
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 239000000284 extract Substances 0.000 claims description 12
- 238000010586 diagram Methods 0.000 description 4
- 239000003814 drug Substances 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a commodity propelling method and device based on image character recognition. The method comprises the following steps of: S1, acquiring a character area in images to be recognized; S2, carrying out character recognition on the character area; S3, inquiring a commodity warehouse based on the recognition result to acquire the commodity information corresponding to the recognition result; and S4, propelling a commodity inquiring list containing the commodity information. By the method and device, the commodity information can be directly acquired by uploading images without manfully searching the commodity information from a great amount of searching results through a search engine, so that the user operation is greatly reduced and convenience is brought.
Description
[ technical field ] A method for producing a semiconductor device
The invention relates to the technical field of computer application, in particular to a commodity pushing method and device based on image character recognition.
[ background of the invention ]
With the rapid development of the mobile internet, the application of the image collected based on the camera of the mobile terminal is more and more extensive. The image character recognition technology recognizes characters in the image and converts the characters into text characters, so that the burden of inputting corresponding character information by a user is reduced, and the user can store and edit the corresponding character information conveniently.
In the actual application process, a user wants to query relevant information of a certain commodity, such as commodity use, manufacturer, price, where the commodity is sold and the like, after seeing the commodity, the existing method is that the user manually inputs a commodity name and the like through a search engine to serve as query, and the desired commodity information is searched from a large number of search results, obviously, the method is very complicated in operation and needs a large amount of manual operation.
[ summary of the invention ]
In view of this, the invention provides a method and a device for pushing a commodity based on image character recognition, so as to reduce operations of a user for obtaining commodity information and achieve more convenience.
The specific technical scheme is as follows:
a commodity pushing method based on image character recognition comprises the following steps:
s1, acquiring a character area in the image to be recognized;
s2, performing character recognition on the character area;
s3, inquiring the commodity library by using the identification result to obtain commodity information corresponding to the identification result;
and S4, pushing a commodity inquiry list containing the commodity information.
According to a preferred embodiment of the present invention, the step S1 specifically includes:
the method comprises the steps that a server receives an image to be identified sent by a mobile terminal, and extracts a character area from the image to be identified; or,
the server receives the character area which is extracted from the image to be identified and sent by the mobile terminal.
According to a preferred embodiment of the present invention, the step S2 specifically includes:
carrying out binarization on the character area;
dividing the binarized character area into character blocks;
extracting the characteristic information of each block, matching the characteristic information with a characteristic database, and taking the matching result as the recognition result of each block;
and combining the recognition results of the character blocks in sequence to obtain the recognition result of the character area.
According to a preferred embodiment of the present invention, the commodity library includes more than one category of commodity libraries;
querying all commodity libraries in the step S3; or,
in the step S1, the personalized option content selected by the user is also obtained, and in the step S3, the commodity library corresponding to the personalized option content selected by the user is queried.
According to a preferred embodiment of the present invention, the querying the commodity library to obtain the commodity information corresponding to the identification result specifically includes:
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and the recognition result, and enabling the commodity information with the character matching weights rearranged in the first n1 to be contained in a commodity query list, wherein n1 is a preset positive integer; or,
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and recognition results, calculating image matching weights of the commodity information according to similarity between the image to be recognized and the images in the commodity information of the commodity library, calculating query weights corresponding to the commodity information by combining the character matching weights and the image matching weights, and containing the commodity information with the query weights rearranged in the first n2 in a commodity query list, wherein n2 is a preset positive integer; or,
calculating a character matching weight of the commodity information according to semantic similarity between character content in the commodity information of the commodity library and an identification result, calculating an image matching weight of the commodity information according to similarity between the image to be identified and the image in the commodity information of the commodity library, calculating a query weight corresponding to the commodity information by combining the character matching weight and the image matching weight, calculating a selection weight of the commodity information based on a queried condition of the commodity information, calculating a total weight of the commodity information by combining the query weight and the selection weight, and generating a commodity query list by arranging the commodity information with the total weight values in the top n3, wherein n3 is a preset positive integer.
According to a preferred embodiment of the present invention, the calculating the selection weight of the commodity information includes:
calculating the selection weight of the commodity information according to the total times of the commodity information which is inquired, wherein the larger the total times, the larger the selection weight value is; or,
determining the commodity weight of the commodity information by using the total inquired times of the commodity information, determining the user personalized weight by using the total times of checking all the commodity information of the category to which the commodity information belongs by the current user when the total times of the commodity weight is larger, and determining the selection weight of the commodity information by using the product of the commodity weight of the commodity information and the user personalized weight.
A commodity pushing device based on image character recognition comprises:
the region acquisition unit is used for acquiring a character region in the image to be recognized;
the character recognition unit is used for carrying out character recognition on the character area;
the commodity inquiry unit is used for inquiring the commodity library by the identification result of the character identification unit to obtain the commodity information corresponding to the identification result;
and the result pushing unit is used for pushing the commodity inquiry list containing the commodity information.
According to a preferred embodiment of the present invention, the area obtaining unit receives an image to be recognized sent by a mobile terminal, and extracts a text area from the image to be recognized; or the receiving mobile terminal extracts and transmits the character area from the image to be recognized.
According to a preferred embodiment of the present invention, the character recognition unit specifically performs: the method comprises the steps of binarizing a character area, dividing the binarized character area into character blocks, extracting characteristic information of the character blocks, matching the characteristic information with a characteristic database, using matching results as recognition results of the character blocks, and combining the recognition results of the character blocks in sequence to obtain recognition results of the character area.
According to a preferred embodiment of the present invention, the commodity library includes more than one category of commodity libraries;
the commodity inquiring unit inquires all commodity libraries; or,
the area acquisition unit also acquires the personalized option content selected by the user, and the commodity query unit queries a commodity library corresponding to the personalized option content selected by the user.
According to a preferred embodiment of the present invention, when the commodity searching unit searches the commodity library to obtain the commodity information corresponding to the identification result, specifically:
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and the recognition result, and enabling the commodity information with the character matching weights rearranged in the first n1 to be contained in a commodity query list, wherein n1 is a preset positive integer; or,
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and recognition results, calculating image matching weights of the commodity information according to similarity between the image to be recognized and the images in the commodity information of the commodity library, calculating query weights corresponding to the commodity information by combining the character matching weights and the image matching weights, and containing the commodity information with the query weights rearranged in the first n2 in a commodity query list, wherein n2 is a preset positive integer; or,
calculating a character matching weight of the commodity information according to semantic similarity between character content in the commodity information of the commodity library and an identification result, calculating an image matching weight of the commodity information according to similarity between the image to be identified and the image in the commodity information of the commodity library, calculating a query weight corresponding to the commodity information by combining the character matching weight and the image matching weight, calculating a selection weight of the commodity information based on a queried condition of the commodity information, calculating a total weight of the commodity information by combining the query weight and the selection weight, and generating a commodity query list by arranging the commodity information with the total weight values in the top n3, wherein n3 is a preset positive integer.
According to a preferred embodiment of the present invention, when the commodity searching unit calculates the selection weight of the commodity information, specifically:
calculating the selection weight of the commodity information according to the total times of the commodity information which is inquired, wherein the larger the total times, the larger the selection weight value is; or,
determining the commodity weight of the commodity information by using the total inquired times of the commodity information, determining the user personalized weight by using the total times of checking all the commodity information of the category to which the commodity information belongs by the current user when the total times of the commodity weight is larger, and determining the selection weight of the commodity information by using the product of the commodity weight of the commodity information and the user personalized weight.
According to the technical scheme, on the basis of image character recognition, the commodity information corresponding to the recognition result is obtained by querying the commodity library through the recognition result, and the commodity query list containing the commodity information is pushed, so that a user can directly obtain the commodity information in an image uploading mode without manually searching the commodity information from a large number of search results through a search engine, the operation of the user is greatly reduced, and the realization is more convenient.
[ description of the drawings ]
Fig. 1 is a flowchart of a commodity pushing method based on image character recognition according to an embodiment of the present invention;
FIG. 2 is a block diagram of a system according to an embodiment of the present invention;
fig. 3 is a structural diagram of a product pushing device based on image character recognition according to an embodiment of the present invention;
fig. 4 and fig. 5 are schematic diagrams illustrating two display effects of the mobile terminal according to the embodiment of the present invention.
[ detailed description ] embodiments
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
The first embodiment,
Fig. 1 is a flowchart of a commodity pushing method based on image character recognition according to an embodiment of the present invention, and as shown in fig. 1, the method may include the following steps:
step 101: and acquiring a character area in the image to be recognized.
The server acquires an image containing text information sent by the mobile terminal, wherein the image can be an original image shot by the mobile terminal, and the server extracts a text area in the image to be recognized in the step. Or the image may be an original image shot by the mobile terminal, and after the character area in the image to be recognized is extracted, the character area in the image to be recognized is sent to the server.
The existing method can be adopted when extracting the text region, and the text region can be extracted after removing the image background, but the method can be adopted, but is not limited to, the following method:
firstly, color run-length coding is carried out according to color Euclidean distances, then color clustering is carried out, character layers are generated and selected based on clustering results, for example, a connected domain with the area larger than a certain value is reserved, each image layer is generated based on the Euclidean distances between the connected domain and each color clustering center, finally, the character layer, the noise layer or the background layer is determined according to the relation between the number of pixels of each image layer and the number of pixels of the segmentation threshold of the layer, and finally, the character layer, namely the character region is obtained after the noise layer and the background layer are taken out.
Selecting a large number of character sample images and images without characters, and extracting edge information of the two images by using a canny operator to serve as training samples of the sparse representation classification dictionary; inputting the two types of training samples into a classification sparse representation dictionary training algorithm to obtain a character sparse representation classification dictionary and a non-character sparse representation classification dictionary; converting an image to be identified into a gray image, and extracting edge information of the gray image by using a canny operator; extracting candidate character regions in the gray level image edge information by utilizing sparse representation based on a classification dictionary; respectively connecting isolated edges of the candidate character areas into larger areas by using a run-length smoothing algorithm in the horizontal direction and the vertical direction, then carrying out projection analysis to find out corresponding character lines, and simultaneously discarding the isolated edges except the character lines in the candidate character areas; and identifying the detected character area.
If the mobile terminal extracts the character area, the existing character area extraction software or a manual mode can be adopted to extract the character area.
In addition, the number of the character areas acquired in this step may be one, or may be two or more. Since the content in this step is the prior art, it is not described herein again.
Step 102: and performing character recognition on the character area.
The process of character recognition for the character area can also adopt the prior art, namely the following steps are included: carrying out binarization on the character area; dividing the binarized character area into character blocks; extracting the characteristic information of each block, matching the characteristic information with the characteristic database, taking the matching result as the recognition result of each block, and combining the recognition results of each block in sequence to obtain the recognition result of the character area.
In addition, the text recognition mode is various, and any other mode capable of realizing text recognition can be adopted besides the above mode, which is not described in detail.
Step 103: and querying the commodity library by using the identification result to obtain commodity information corresponding to the identification result.
The commodity library queried in this step may be an entity commodity library or a virtual commodity library, and these commodity libraries may be local commodity libraries, network commodity libraries, or commodity libraries in which an access interface is opened by a third party.
The physical commodity library may include, but is not limited to, various physical commodity libraries such as a book commodity library including book commodity information, a food commodity library including food commodity information, a clothing commodity library including clothing commodity information, and a medicine commodity library including medicine information. The virtual goods library may include, but is not limited to, an electronic book goods library including electronic book goods information, a game point card goods library including game point card goods information, an application software goods library including application software goods information, a service goods library including service goods information, and other virtual goods libraries.
When the entity commodity library or the virtual commodity library is inquired, the character matching weight of the character content and the recognition result in the commodity information is calculated, the character matching weight depends on the semantic similarity between the commodity information and the recognition result, and the commodity information with the character matching weight reaching a preset character matching weight threshold value is used as the inquiry result.
The semantic similarity determination method can adopt the prior art, aims to calculate the semantic similarity of the text content of the commodity information and the recognition result, and can adopt, but is not limited to, the following modes: extracting key semantic words in the recognition result, inquiring the character content of the commodity information, obtaining the number of successfully matched key semantic words, and taking the number as the determination basis of semantic similarity. The larger the number of successfully matched key semantic vocabulary words is, the larger the semantic similarity is, and the larger the corresponding word matching weight is.
In addition, in view of the particularity of the physical commodity, the physical commodity may have an image, and when the physical commodity library is queried, an image matching weight between the image to be recognized and the image in the commodity information of the commodity library may be further calculated, where the image matching weight depends on the similarity between the image in the commodity information and the image to be recognized. And then calculating the query weight corresponding to the commodity information by combining the character matching weight and the image matching weight corresponding to the commodity information. When the query weight corresponding to the commodity information is calculated, the product of the character matching weight and the image matching weight or the sum of the character matching weight and the image matching weight can be used as the corresponding query weight.
The similarity between the image in the commodity information and the image to be identified may also be determined by using the prior art, and may be determined by using, but not limited to, the method: respectively extracting color histograms of the image to be identified and the commodity image, calculating Euclidean distance between the color histograms, and determining the similarity between the two images based on the Euclidean distance. The smaller the Euclidean distance is, the greater the similarity between two images is, and the greater the corresponding image matching weight is.
In one implementation, after all the commodity libraries are queried, the commodity information with the letter matching weight rearranged in the top n1 is included in the commodity query list for returning to the mobile terminal. Specifically, the commodity information in which the query right is rearranged to the top n2 commodities information may be included in the commodity query list for return to the mobile terminal for the commodity information of the physical commodity. Wherein n1 and n2 are preset positive integers.
Another implementation manner is that the mobile terminal provides personalized setting options for the user, the content of the options selected by the user is simultaneously sent when the image is sent to the server, and the server only queries the commodity library of the category corresponding to the content of the options selected by the user when querying the commodity libraries of the categories in the step. Then the information of the commodities with the letter matching weight rearranged in the top n1 is contained in the commodity inquiry list for returning to the mobile terminal. Specifically, the commodity information in which the query right is rearranged to the top n2 commodities information may be included in the commodity query list for return to the mobile terminal for the commodity information of the physical commodity. Wherein n1 and n2 are preset positive integers.
For example, the mobile terminal provides personalized setting options such as physical goods, virtual goods, and the like, or more specifically personalized setting options such as book goods, food goods, clothing goods, electronic book goods, game click goods, application software goods, service goods, and the like, to the user, if the user takes an image of a food package through the mobile terminal, the user may select the option of the food goods, and then the mobile terminal sends the image and the content of the option selected by the user to the server, and when the server queries the goods library according to the text recognition result of the image, the server may query only the food goods library, generate a goods query list according to the obtained query result, and then return the goods query list to the mobile terminal in step 104. Of course, the user may also select more than one option.
When the commodity library is inquired, all the commodity libraries or the commodity libraries of the types corresponding to the option contents selected by the user are inquired, but when the commodity information is returned, the total weight of each commodity information is calculated by combining the inquiry weight and the selection weight of the commodity information, the commodity information with the first n3 total weight values is generated into a commodity inquiry list for returning to the mobile terminal, and n3 is a preset positive integer.
The selection weight of the commodity information can be determined by the following methods without limitation: first, the total number of times the commodity information is queried is greater, the greater the total number of times, the greater the weight value is, and the total number of times is referred to as the total number of times queried by all users. Secondly, determining the commodity weight corresponding to the commodity information by using the total times of querying the commodity information by all users, determining the user personalized weight by using the total times of viewing all the commodity information of the category to which the commodity information belongs by the current user (i.e. pushing the commodity information to the mobile terminal, the user can view the commodity information of some categories, such as pushing the food commodity information and clothing commodity information to the user, and if the user views the food commodity information, updating the times of viewing the food commodity information for updating the user personalized weight of the food commodity information), and determining the selection weight of the commodity information by using the product of the commodity weight of the commodity information and the user personalized weight.
After the selection weight of the commodity information is obtained, the total weight of the commodity information may be obtained by using the product of the selection weight of the commodity information and the query weight, or the total weight of the commodity information may be obtained by summing or the like.
In addition, the step can be based on all the character information of the recognition result when the commodity library is inquired, and also can be based on the key meaning character information obtained after the word segmentation is carried out on the recognition result.
Step 104: and pushing a commodity inquiry list containing corresponding commodity information to the mobile terminal.
After the server returns the commodity information to the mobile terminal, the user can acquire the corresponding commodity information from the display of the mobile terminal. And the commodity information may be in more than one category, if the user checks one or more categories, the commodity information is reported to the server, the total times of inquiring all the users of the commodity information are updated by the server, and meanwhile, the corresponding selection of the commodity category to which the commodity information belongs is updated.
In addition, in addition to returning the goods inquiry list to the mobile terminal, the recognition result may be simultaneously returned to the mobile terminal.
The method provided by the invention is described above, and the device provided by the invention is described in detail by the embodiment below. For convenience of understanding, a system applied by the above method of the present invention is first described, and as shown in fig. 2, the system is composed of a mobile terminal and a server, wherein the mobile terminal can send a shot image containing characters to the server as an image to be recognized, and extract a character region from the image, or the mobile terminal extracts a character region from the shot image containing characters as an image to be recognized, and sends the character region to the server. And then the server executes the flow shown in the first embodiment and returns the commodity inquiry list to the mobile terminal. The device provided by the following second embodiment of the invention is arranged in the server and used for completing the flow shown in the first embodiment.
Example II,
Fig. 3 is a structural diagram of an image character recognition apparatus according to a second embodiment of the present invention, as shown in fig. 3, the apparatus includes: an area acquisition unit 301, a character recognition unit 302, a product inquiry unit 303 and a result pushing unit 304.
First, the region acquisition unit 301 acquires a text region in an image to be recognized.
Here, the area acquisition unit 301 receives an image to be recognized sent by the mobile terminal, and extracts a text area from the image to be recognized; or the receiving mobile terminal extracts and transmits the character area from the image to be recognized. When extracting the text area, two ways described in step 101 in the first embodiment can be adopted, and since the contents of this part are prior art, detailed description is omitted here.
The character recognition unit 302 then performs character recognition on the character area. The specific identification process may include: the method comprises the steps of binarizing a character area, dividing the binarized character area into character blocks, extracting characteristic information of the character blocks, matching the characteristic information with a characteristic database, using matching results as recognition results of the character blocks, and combining the recognition results of the character blocks in sequence to obtain recognition results of the character area.
The product searching unit 303 searches the product library using the recognition result of the character recognition unit 302 to obtain product information corresponding to the recognition result.
Since the commodity library according to the present invention includes one or more types of commodity libraries, that is, may be a single type of commodity library or may be a plurality of types of commodity libraries, the commodity searching unit 303 may execute any of the following when searching for a commodity library:
the commodity inquiry unit inquires all commodity libraries; or,
the area acquisition unit also acquires the personalized option content selected by the user, and the commodity inquiry unit inquires a commodity library corresponding to the personalized option content selected by the user.
When querying the commodity library to obtain the commodity information corresponding to the identification result, the following implementation mode can be adopted:
the first embodiment: and calculating the character matching weight of the commodity information according to the semantic similarity between the character content in the commodity information of the commodity library and the recognition result, and rearranging the character matching weight to the first n1 commodity information to be contained in a commodity query list, wherein n1 is a preset positive integer.
The second embodiment: calculating the character matching weight of the commodity information according to the semantic similarity between the character content in the commodity information of the commodity library and the recognition result, calculating the image matching weight of the commodity information according to the similarity between the image to be recognized and the image in the commodity information of the commodity library, calculating the query weight corresponding to the commodity information by combining the character matching weight and the image matching weight, and containing the commodity information with the query weight rearranged in the top n2 in a commodity query list, wherein n2 is a preset positive integer.
Third embodiment: calculating a character matching weight of the commodity information according to semantic similarity between character content in the commodity information of the commodity library and an identification result, calculating an image matching weight of the commodity information according to similarity between the image to be identified and the image in the commodity information of the commodity library, calculating a query weight corresponding to the commodity information by combining the character matching weight and the image matching weight, calculating a selection weight of the commodity information based on a queried condition of the commodity information, calculating a total weight of the commodity information by combining the query weight and the selection weight, and generating a commodity query list by arranging the commodity information with the total weight values in the top n3, wherein n3 is a preset positive integer.
Specifically, when the product querying unit 303 calculates the selection weight of the product information, the following method may be specifically adopted:
the method comprises the steps of calculating the selection weight of the commodity information according to the total times of querying the commodity information by all users, wherein the larger the total times is, the larger the selection weight value is.
And secondly, determining the commodity weight of the commodity information by using the total inquired times of the commodity information, wherein the larger the total times is, the larger the commodity weight is, determining the user personalized weight by using the total times of checking all the commodity information of the category to which the commodity information belongs by the current user, and determining the selection weight of the commodity information by using the product of the commodity weight of the commodity information and the user personalized weight.
Finally, the result pushing unit 304 pushes the commodity query list containing the commodity information. The recognition result can also be returned to the mobile terminal at the same time.
After the commodity inquiry list is returned to the mobile terminal, the user can acquire commodity information from the display of the mobile terminal. Moreover, the commodity information may be in more than one category, and if the user views one or more categories, the commodity information is reported to the server, the commodity querying unit 303 updates the total number of times that each commodity information is queried by all users, and updates the selection corresponding to the commodity category to which the commodity information belongs.
By the method and the device, the user can obtain the corresponding commodity information in a picture uploading mode without manually obtaining the commodity information from a large number of search results in a search engine mode, and the method and the device are obviously more convenient and labor-saving.
For example, after the user takes an image containing the text "thousand years exclamation" through the mobile terminal, and sends the image to the server, the server performs image and text recognition and queries the commodity library through the above processes, and returns a recognition result and a commodity query list, where the display mode of the commodity query list is not limited in the present invention, and may adopt any mode, such as the mode of the reference box shown in fig. 4.
For another example, after the user shoots an image containing the characters "dark sky please close eyes" through the mobile terminal and sends the image to the server, the server performs image character recognition and queries the commodity library through the above processes, and the returned commodity query list contains various categories of commodities, wherein the display mode of each category of commodities is not limited in the present invention, for example, the mode of adopting labels shown in fig. 5.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (12)
1. A commodity pushing method based on image character recognition is characterized by comprising the following steps:
s1, acquiring a character area in the image to be recognized;
s2, performing character recognition on the character area;
s3, inquiring the commodity library by using the identification result to obtain commodity information corresponding to the identification result;
and S4, pushing a commodity inquiry list containing the commodity information.
2. The method according to claim 1, wherein the step S1 specifically includes:
the method comprises the steps that a server receives an image to be identified sent by a mobile terminal, and extracts a character area from the image to be identified; or,
the server receives the character area which is extracted from the image to be identified and sent by the mobile terminal.
3. The method according to claim 1, wherein the step S2 specifically includes:
carrying out binarization on the character area;
dividing the binarized character area into character blocks;
extracting the characteristic information of each block, matching the characteristic information with a characteristic database, and taking the matching result as the recognition result of each block;
and combining the recognition results of the character blocks in sequence to obtain the recognition result of the character area.
4. The method of claim 1, wherein the commodity library comprises more than one category of commodity library;
querying all commodity libraries in the step S3; or,
in the step S1, the personalized option content selected by the user is also obtained, and in the step S3, the commodity library corresponding to the personalized option content selected by the user is queried.
5. The method according to claim 4, wherein the querying the commodity library to obtain the commodity information corresponding to the identification result specifically comprises:
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and the recognition result, and enabling the commodity information with the character matching weights rearranged in the first n1 to be contained in a commodity query list, wherein n1 is a preset positive integer; or,
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and recognition results, calculating image matching weights of the commodity information according to similarity between the image to be recognized and the images in the commodity information of the commodity library, calculating query weights corresponding to the commodity information by combining the character matching weights and the image matching weights, and containing the commodity information with the query weights rearranged in the first n2 in a commodity query list, wherein n2 is a preset positive integer; or,
calculating a character matching weight of the commodity information according to semantic similarity between character content in the commodity information of the commodity library and an identification result, calculating an image matching weight of the commodity information according to similarity between the image to be identified and the image in the commodity information of the commodity library, calculating a query weight corresponding to the commodity information by combining the character matching weight and the image matching weight, calculating a selection weight of the commodity information based on a queried condition of the commodity information, calculating a total weight of the commodity information by combining the query weight and the selection weight, and generating a commodity query list by arranging the commodity information with the total weight values in the top n3, wherein n3 is a preset positive integer.
6. The method of claim 5, wherein the calculating the selection weight of the merchandise information comprises:
calculating the selection weight of the commodity information according to the total times of the commodity information which is inquired, wherein the larger the total times, the larger the selection weight value is; or,
determining the commodity weight of the commodity information by using the total inquired times of the commodity information, determining the user personalized weight by using the total times of checking all the commodity information of the category to which the commodity information belongs by the current user when the total times of the commodity weight is larger, and determining the selection weight of the commodity information by using the product of the commodity weight of the commodity information and the user personalized weight.
7. A commodity pushing device based on image character recognition is characterized by comprising:
the region acquisition unit is used for acquiring a character region in the image to be recognized;
the character recognition unit is used for carrying out character recognition on the character area;
the commodity inquiry unit is used for inquiring the commodity library by the identification result of the character identification unit to obtain the commodity information corresponding to the identification result;
and the result pushing unit is used for pushing the commodity inquiry list containing the commodity information.
8. The device according to claim 7, wherein the area acquisition unit receives an image to be recognized sent by a mobile terminal, and extracts a text area from the image to be recognized; or the receiving mobile terminal extracts and transmits the character area from the image to be recognized.
9. The apparatus of claim 7, wherein the text recognition unit specifically performs: the method comprises the steps of binarizing a character area, dividing the binarized character area into character blocks, extracting characteristic information of the character blocks, matching the characteristic information with a characteristic database, using matching results as recognition results of the character blocks, and combining the recognition results of the character blocks in sequence to obtain recognition results of the character area.
10. The apparatus of claim 7, wherein the commodity library comprises more than one category of commodity library;
the commodity inquiring unit inquires all commodity libraries; or,
the area acquisition unit also acquires the personalized option content selected by the user, and the commodity query unit queries a commodity library corresponding to the personalized option content selected by the user.
11. The apparatus according to claim 10, wherein when the product querying unit queries the product library to obtain the product information corresponding to the identification result, specifically:
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and the recognition result, and enabling the commodity information with the character matching weights rearranged in the first n1 to be contained in a commodity query list, wherein n1 is a preset positive integer; or,
calculating character matching weights of the commodity information according to semantic similarity between character contents in the commodity information of the commodity library and recognition results, calculating image matching weights of the commodity information according to similarity between the image to be recognized and the images in the commodity information of the commodity library, calculating query weights corresponding to the commodity information by combining the character matching weights and the image matching weights, and containing the commodity information with the query weights rearranged in the first n2 in a commodity query list, wherein n2 is a preset positive integer; or,
calculating a character matching weight of the commodity information according to semantic similarity between character content in the commodity information of the commodity library and an identification result, calculating an image matching weight of the commodity information according to similarity between the image to be identified and the image in the commodity information of the commodity library, calculating a query weight corresponding to the commodity information by combining the character matching weight and the image matching weight, calculating a selection weight of the commodity information based on a queried condition of the commodity information, calculating a total weight of the commodity information by combining the query weight and the selection weight, and generating a commodity query list by arranging the commodity information with the total weight values in the top n3, wherein n3 is a preset positive integer.
12. The apparatus according to claim 11, wherein the product search unit, when calculating the selection weight of the product information, specifically:
calculating the selection weight of the commodity information according to the total times of the commodity information which is inquired, wherein the larger the total times, the larger the selection weight value is; or,
determining the commodity weight of the commodity information by using the total inquired times of the commodity information, determining the user personalized weight by using the total times of checking all the commodity information of the category to which the commodity information belongs by the current user when the total times of the commodity weight is larger, and determining the selection weight of the commodity information by using the product of the commodity weight of the commodity information and the user personalized weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012102793672A CN102831200A (en) | 2012-08-07 | 2012-08-07 | Commodity propelling method and device based on image character recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012102793672A CN102831200A (en) | 2012-08-07 | 2012-08-07 | Commodity propelling method and device based on image character recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102831200A true CN102831200A (en) | 2012-12-19 |
Family
ID=47334337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012102793672A Pending CN102831200A (en) | 2012-08-07 | 2012-08-07 | Commodity propelling method and device based on image character recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102831200A (en) |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103064936A (en) * | 2012-12-24 | 2013-04-24 | 北京百度网讯科技有限公司 | Voice-input-based image information extraction analysis 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 |
| CN103412938A (en) * | 2013-08-22 | 2013-11-27 | 成都数之联科技有限公司 | Commodity price comparing method based on picture interactive type multiple-target extraction |
| CN104008388A (en) * | 2014-06-06 | 2014-08-27 | 杨军辉 | Method and system for obtaining merchant business data by recognizing product identification |
| CN104240096A (en) * | 2014-08-27 | 2014-12-24 | 小米科技有限责任公司 | Information display method and device and electronic equipment |
| CN104268168A (en) * | 2014-09-10 | 2015-01-07 | 百度在线网络技术（北京）有限公司 | Method and device for pushing information to user |
| CN105095446A (en) * | 2015-07-24 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Medicine search processing method, server and terminal device |
| CN105321146A (en) * | 2015-09-25 | 2016-02-10 | 广东小天才科技有限公司 | Method and device for processing topic picture shot by mobile terminal |
| 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 |
| 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 |
| CN105912642A (en) * | 2016-04-08 | 2016-08-31 | 世纪禾光科技发展（北京）有限公司 | Product price data acquisition method and system |
| CN106126755A (en) * | 2016-08-24 | 2016-11-16 | 广东华邦云计算股份有限公司 | A kind of purchase method based on image recognition |
| CN106294527A (en) * | 2015-06-26 | 2017-01-04 | 阿里巴巴集团控股有限公司 | A kind of information recommendation method and equipment |
| CN106708823A (en) * | 2015-07-20 | 2017-05-24 | 阿里巴巴集团控股有限公司 | Search processing method, apparatus and system |
| CN107292642A (en) * | 2016-03-31 | 2017-10-24 | 苏宁云商集团股份有限公司 | A kind of Method of Commodity Recommendation and system based on image |
| 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 |
| CN107330391A (en) * | 2017-06-26 | 2017-11-07 | 北京小米移动软件有限公司 | Product information reminding method and device |
| CN107580047A (en) * | 2017-08-31 | 2018-01-12 | 广东美的制冷设备有限公司 | Device pushing method, electronic device and computer readable storage medium |
| CN107798070A (en) * | 2017-09-26 | 2018-03-13 | 平安普惠企业管理有限公司 | A kind of web data acquisition methods and terminal device |
| CN108804978A (en) * | 2017-04-28 | 2018-11-13 | 腾讯科技（深圳）有限公司 | A kind of printed page analysis method and device |
| CN109429077A (en) * | 2017-08-24 | 2019-03-05 | 北京搜狗科技发展有限公司 | Method for processing video frequency and device, for the device of video processing |
| CN110858233A (en) * | 2018-08-17 | 2020-03-03 | 珠海格力电器股份有限公司 | Mobile terminal information recommendation system and method based on image understanding |
| CN111258409A (en) * | 2020-05-06 | 2020-06-09 | 北京深光科技有限公司 | Feature point identification method and device for man-machine interaction |
| CN112801737A (en) * | 2021-01-27 | 2021-05-14 | 广州微框物联科技有限公司 | Lipstick shopping guide method and device of intelligent mirror |
| CN113724030A (en) * | 2020-07-23 | 2021-11-30 | 景德镇陶瓷大学 | Method and system for customizing articles on line |
Citations (6)
| 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 |
| CN101136096A (en) * | 2006-08-31 | 2008-03-05 | 林�智 | Dress ornament evaluating and matching system and method on internet |
| US20080059526A1 (en) * | 2006-09-01 | 2008-03-06 | Sony Corporation | Playback apparatus, searching method, and program |
| CN101414307A (en) * | 2008-11-26 | 2009-04-22 | 阿里巴巴集团控股有限公司 | Method and server for providing picture searching |
| 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 |
-
2012
- 2012-08-07 CN CN2012102793672A patent/CN102831200A/en active Pending
Patent Citations (6)
| 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 |
| CN101136096A (en) * | 2006-08-31 | 2008-03-05 | 林�智 | Dress ornament evaluating and matching system and method on internet |
| 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 |
| CN101414307A (en) * | 2008-11-26 | 2009-04-22 | 阿里巴巴集团控股有限公司 | Method and server for providing picture searching |
| 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 |
Non-Patent Citations (1)
| Title |
|---|
| 唐一之: "《无形的市场——知识本体与网络消费研究（2009年1月第1版）》", 31 January 2009, 湖南师范大学出版社 * |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103064936B (en) * | 2012-12-24 | 2018-03-30 | 北京百度网讯科技有限公司 | A kind of image information extraction and analytical method and device based on phonetic entry |
| CN103064936A (en) * | 2012-12-24 | 2013-04-24 | 北京百度网讯科技有限公司 | Voice-input-based image information extraction analysis 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 |
| CN103412938A (en) * | 2013-08-22 | 2013-11-27 | 成都数之联科技有限公司 | Commodity price comparing method based on picture interactive type multiple-target extraction |
| CN103412938B (en) * | 2013-08-22 | 2016-06-29 | 成都数之联科技有限公司 | A kind of commodity price-comparing method extracted based on picture interactive multiobjective |
| CN104008388A (en) * | 2014-06-06 | 2014-08-27 | 杨军辉 | Method and system for obtaining merchant business data by recognizing product identification |
| CN104240096A (en) * | 2014-08-27 | 2014-12-24 | 小米科技有限责任公司 | Information display method and device and electronic equipment |
| 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 |
| CN104268168A (en) * | 2014-09-10 | 2015-01-07 | 百度在线网络技术（北京）有限公司 | Method and device for pushing information to user |
| CN106294527A (en) * | 2015-06-26 | 2017-01-04 | 阿里巴巴集团控股有限公司 | A kind of information recommendation method and equipment |
| CN106708823A (en) * | 2015-07-20 | 2017-05-24 | 阿里巴巴集团控股有限公司 | Search processing method, apparatus and system |
| CN105095446A (en) * | 2015-07-24 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Medicine search processing method, server and terminal device |
| CN105321146A (en) * | 2015-09-25 | 2016-02-10 | 广东小天才科技有限公司 | Method and device for processing topic picture shot by mobile terminal |
| 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 |
| CN107292642A (en) * | 2016-03-31 | 2017-10-24 | 苏宁云商集团股份有限公司 | A kind of Method of Commodity Recommendation and system based on image |
| CN105912642A (en) * | 2016-04-08 | 2016-08-31 | 世纪禾光科技发展（北京）有限公司 | Product price data acquisition method and system |
| CN106126755A (en) * | 2016-08-24 | 2016-11-16 | 广东华邦云计算股份有限公司 | A kind of purchase method based on image recognition |
| CN108804978B (en) * | 2017-04-28 | 2022-04-12 | 腾讯科技（深圳）有限公司 | Layout analysis method and device |
| CN108804978A (en) * | 2017-04-28 | 2018-11-13 | 腾讯科技（深圳）有限公司 | A kind of printed page analysis method and device |
| 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 |
| CN107330391A (en) * | 2017-06-26 | 2017-11-07 | 北京小米移动软件有限公司 | Product information reminding method and device |
| CN109429077A (en) * | 2017-08-24 | 2019-03-05 | 北京搜狗科技发展有限公司 | Method for processing video frequency and device, for the device of video processing |
| CN107580047A (en) * | 2017-08-31 | 2018-01-12 | 广东美的制冷设备有限公司 | Device pushing method, electronic device and computer readable storage medium |
| WO2019041595A1 (en) * | 2017-08-31 | 2019-03-07 | 广东美的制冷设备有限公司 | Device pushing method, electronic device, and computer readable storage medium |
| CN107798070A (en) * | 2017-09-26 | 2018-03-13 | 平安普惠企业管理有限公司 | A kind of web data acquisition methods and terminal device |
| CN110858233A (en) * | 2018-08-17 | 2020-03-03 | 珠海格力电器股份有限公司 | Mobile terminal information recommendation system and method based on image understanding |
| CN111258409A (en) * | 2020-05-06 | 2020-06-09 | 北京深光科技有限公司 | Feature point identification method and device for man-machine interaction |
| CN113724030A (en) * | 2020-07-23 | 2021-11-30 | 景德镇陶瓷大学 | Method and system for customizing articles on line |
| CN112801737A (en) * | 2021-01-27 | 2021-05-14 | 广州微框物联科技有限公司 | Lipstick shopping guide method and device of intelligent mirror |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102831200A (en) | Commodity propelling method and device based on image character recognition | |
| CN102855480A (en) | Method and device for recognizing characters in image | |
| US9042659B2 (en) | Method and system for fast and robust identification of specific product images | |
| CN111400607B (en) | Search content output method and device, computer equipment and readable storage medium | |
| CN110110577B (en) | Method and device for identifying dish name, storage medium and electronic device | |
| WO2020005731A1 (en) | Text entity detection and recognition from images | |
| KR20130142191A (en) | Robust feature matching for visual search | |
| US8254678B2 (en) | Image segmentation | |
| KR20120001285A (en) | Product classification search and shopping information provision service method, server and system through object recognition | |
| CN108492160A (en) | Information recommendation method and device | |
| CN112738556A (en) | Video processing method and device | |
| WO2006122164A2 (en) | System and method for enabling the use of captured images through recognition | |
| CN111767420B (en) | Method and device for generating clothing collocation data | |
| CN110135769A (en) | Kinds of goods attribute fill method and device, storage medium and electric terminal | |
| CN104142955A (en) | A method and terminal for recommending learning courses | |
| US9910864B2 (en) | Method for object recognition, corresponding system, apparatus and computer program product | |
| CN110674388A (en) | Mapping method and device for push item, storage medium and terminal equipment | |
| CN115618109A (en) | Content recommendation method and device, electronic equipment and computer-readable storage medium | |
| CN113657273B (en) | Method, device, electronic equipment and medium for determining commodity information | |
| CN111177450A (en) | Image retrieval cloud identification method and system and computer readable storage medium | |
| JPH11250106A (en) | Automatic search method of registered trademark using content-based video information | |
| CN111460888B (en) | Article identification method and device based on machine learning | |
| CN113705209A (en) | Subtitle generating method and device, electronic equipment and storage medium | |
| CN114117110A (en) | Commodity data processing method and device, storage medium and processor | |
| Hwang et al. | Enabling product recognition and tracking based on text detection for mobile augmented reality |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20121219 |