CN109902733A - The method, apparatus and storage medium of typing Item Information - Google Patents
The method, apparatus and storage medium of typing Item Information Download PDFInfo
- Publication number
- CN109902733A CN109902733A CN201910132029.8A CN201910132029A CN109902733A CN 109902733 A CN109902733 A CN 109902733A CN 201910132029 A CN201910132029 A CN 201910132029A CN 109902733 A CN109902733 A CN 109902733A
- Authority
- CN
- China
- Prior art keywords
- item information
- article
- information
- logged
- item
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The application provides the method, apparatus and storage medium of a kind of typing Item Information, and method includes: to obtain the image information of article to be logged;Described image information is identified based on Item Information identification model trained in advance, obtains the matched Item Information of article to be logged;Typing is carried out according to the Item Information.Automatic input Item Information may be implemented in the application, the efficiency of Item Information typing can be improved, and can promote the accuracy of typing Item Information.
Description
Technical field
This application involves technical field of data processing more particularly to a kind of method, apparatus and storage of typing Item Information
Medium.
Background technique
As internet is in the universal of commercial field, more and more businessmans begin to use various article transaction systems to take
Generation tradition relies on the artificial mode for carrying out article trading.
Currently, businessman mostly uses greatly manual or speech recognition typing mode by the title of article, type and other items information
Typing article transaction system.However, above two typing mode is not good solution, not only consumed needed for Input Process
The time taken is longer, and the accuracy of typing Item Information is lower.
Summary of the invention
In view of this, the application provides the method, apparatus and storage medium of a kind of typing Item Information, it is existing to solve
The problem of technical solution of typing Item Information.
Specifically, the application is achieved by the following technical solution:
According to a first aspect of the present application, a kind of method of typing Item Information is proposed, comprising:
Obtain the image information of article to be logged;
Described image information is identified based on Item Information identification model trained in advance, obtains the object to be logged
The matched Item Information of product;
Typing is carried out according to the Item Information.
In one embodiment, described that described image information is known based on Item Information identification model trained in advance
Not, the matched Item Information of article to be logged is obtained, comprising:
By described image information input into Item Information identification model trained in advance, identified based on the Item Information
The output result of model determines the probability of the article to be logged Yu each alternative article information matches;
The matched Item Information of article to be logged according to the determine the probability.
In one embodiment, which is characterized in that the matched article of article to be logged according to the determine the probability
Information, comprising:
The probability is ranked up according to the size of probability value;
The matched Item Information of article to be logged is determined based on the result of the sequence.
In one embodiment, the result based on the sequence determines the matched Item Information of article to be logged,
Include:
The maximum alternative article information of probability value is determined based on the result of the sequence;
The maximum alternative article information of the probability value is determined as the matched Item Information of article to be logged.
In one embodiment, the result based on the sequence determines the matched Item Information of article to be logged,
Include:
Result based on the sequence exports multiple alternative article information and selects for user;
The alternative article information that the user selects is determined as the matched Item Information of article to be logged.
In one embodiment, the method also includes being in advance based on the following steps training Item Information identification model:
Obtain multiple sample item image informations;
Demarcate the corresponding Item Information of each sample item image information;
Using the multiple sample item image information and corresponding Item Information as training set, training Item Information identification
Model.
In one embodiment, described using the multiple sample item image information and corresponding Item Information as training
Collection, training Item Information identification model, comprising:
The deep neural network model based on convolutional neural networks CNN is constructed, and is set for the deep neural network model
A loss function is set, the deep neural network model has n output, and n is the total quantity of Item Information;
Using the multiple sample item image information and corresponding Item Information as training set, to the depth nerve net
Network model carries out successive ignition training, until the numerical value of the loss function is less than preset threshold.
In one embodiment, the output result based on the Item Information identification model determines the article to be logged
With the probability of each alternative article information matches, comprising:
The output result of the Item Information identification model is normalized;
Depth mind in the neural network model is enabled in result after determining the normalized based on argmax function
The maximum result of output result through network;
By the output result for enabling deep neural network it is maximum as a result, be determined as the article to be logged with it is each standby
Select the matched probability of Item Information.
According to a second aspect of the present application, a kind of device of typing Item Information is proposed, comprising:
Image information acquisition module, for obtaining the image information of article to be logged;
Item Information obtains module, for being carried out based on Item Information identification model trained in advance to described image information
Identification, obtains the matched Item Information of article to be logged;
Item Information recording module, for carrying out typing according to the Item Information.
In one embodiment, the Item Information obtains module, comprising:
Informational probability acquiring unit, for by described image information input to Item Information identification model trained in advance
In, the article to be logged and each alternative article information matches are determined based on the output result of the Item Information identification model
Probability;
Item Information obtaining unit is used for the matched Item Information of article to be logged according to the determine the probability.
In one embodiment, the Item Information obtaining unit is also used to:
The probability is ranked up according to the size of probability value;
The matched Item Information of article to be logged is determined based on the result of the sequence.
In one embodiment, the Item Information obtaining unit is also used to:
The maximum alternative article information of probability value is determined based on the result of the sequence;
The maximum alternative article information of the probability value is determined as the matched Item Information of article to be logged.
In one embodiment, the Item Information obtaining unit is also used to:
Result based on the sequence exports multiple alternative article information and selects for user;
The alternative article information that the user selects is determined as the matched Item Information of article to be logged.
According to the third aspect of the application, a kind of computer readable storage medium is proposed, the storage medium is stored with
Computer program, the method that the computer program is used to execute any of the above-described typing Item Information.
The application passes through the image information for obtaining article to be logged, and based on Item Information identification model pair trained in advance
Described image information identified, obtains the matched Item Information of article to be logged, so according to the Item Information into
Row typing due to the image information automatic identification Item Information based on article to be logged and carries out typing, without manual side
Formula or voice mode typing, thus can reduce the entry time of Item Information, improve the efficiency of typing Item Information, and can be with
Promote the accuracy of typing Item Information.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the method for typing Item Information shown in one exemplary embodiment of the application;
Fig. 2 is the process for how obtaining the matched Item Information of article to be logged shown in one exemplary embodiment of the application
Figure;
Fig. 3 is how based on the result of the sequence to determine the object to be logged shown in one exemplary embodiment of the application
The flow chart of the matched Item Information of product;
Fig. 4 be shown in the application another exemplary embodiment how based on the sequence result determine it is described to be logged
The flow chart of the matched Item Information of article;
Fig. 5 is the process of the method for how training Item Information identification model shown in one exemplary embodiment of the application
Figure;
Fig. 6 is a kind of structural block diagram of the device of typing Item Information shown in one exemplary embodiment of the application;
Fig. 7 is a kind of structural block diagram of the device of typing Item Information shown in the application another exemplary embodiment.
Specific embodiment
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistent with the application.On the contrary, they be only with it is such as appended
The example of the consistent device and method of some aspects be described in detail in claims, the application.
It is only to be not intended to be limiting the application merely for for the purpose of describing particular embodiments in term used in this application.
It is also intended in the application and the "an" of singular used in the attached claims, " described " and "the" including majority
Form, unless the context clearly indicates other meaning.It is also understood that term "and/or" used herein refers to and wraps
It may be combined containing one or more associated any or all of project listed.
It will be appreciated that though various information, but this may be described using term first, second, third, etc. in the application
A little information should not necessarily be limited by these terms.These terms are only used to for same type of information being distinguished from each other out.For example, not departing from
In the case where the application range, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as
One information.Depending on context, word as used in this " if " can be construed to " ... when " or " when ...
When " or " in response to determination ".
As internet is in the universal of commercial field, more and more businessmans begin to use various article transaction systems to take
Generation tradition relies on the artificial mode for carrying out article trading.
Currently, businessman mostly uses greatly manual or speech recognition typing mode by the title of article, type and other items information
Typing article transaction system.However, the input speed for being manually entered mode is slower, efficiency of inputting is lower, and is easy error;And language
Sound identifies that the identification accuracy of typing mode is low, and needs artificial recording, thus efficiency of inputting is lower.It can be seen that both
Typing mode is not good solution, is unable to satisfy businessman for the demand of Item Information typing accuracy and efficiency.
In view of this, the application provides the method, apparatus and storage medium of a kind of typing Item Information, it is existing to solve
The problem of technical solution of typing Item Information.
Fig. 1 is a kind of flow chart of the method for typing Item Information shown in the first exemplary embodiment of the application;The reality
Applying example can be used for the terminal device (for example, mobile phone, tablet computer, cashier terminal etc.) of typing Item Information.Such as Fig. 1 institute
Show, the method comprising the steps of S101-S103:
In step s101: obtaining the image information of article to be logged.
In one embodiment, when businessman needs to typing article, the image acquiring device that can use terminal device is obtained
Take the image information of article to be logged.
In one embodiment, above-mentioned image acquiring device can be the devices such as the camera on above-mentioned terminal device, can also
Think the image acquiring device on other terminal devices.It wherein, can be preparatory between above-mentioned terminal device and other terminal devices
Wired or wireless communication connection is established, and then the image information for obtaining article to be logged can be fetched by the communication link.
It is shot to obtain it is worth noting that the image information of above-mentioned article to be logged can treat typing article
Original image information, alternatively, can be treat typing article original image information pre-processed after obtained image information.
It is worth noting that above-mentioned pretreated mode can be chosen according to actual needs by developer, such as will
Original image processing is the image information etc. with predetermined width, height and pixel format, and the present embodiment is to this without limiting.
In step s 102: described image information being identified based on Item Information identification model trained in advance, is obtained
To the matched Item Information of article to be logged.
It in one embodiment, can be based on Item Information trained in advance after obtaining the image information of article to be logged
Identification model identifies described image information, obtains the matched Item Information of article to be logged.
In one embodiment, the particular content of above-mentioned Item Information can be carried out by developer according to actual business requirement
Setting, be such as set as the title of article, number, classification, attribute, price, in feature at least one of, the present embodiment to this not into
Row limits.
In one embodiment, the above-mentioned Item Information identification model of sample image training of acquisition can be advanced with, in turn
It can be identified based on image information of the trained Item Information identification model to the article to be logged of acquisition, to be somebody's turn to do
The matched Item Information of article to be logged.
In one embodiment, the type of above-mentioned Item Information identification model can be carried out according to actual needs by developer
It chooses, such as chooses the deep neural network model based on CNN, the present embodiment is to this without limiting.
In one embodiment, the training method of above-mentioned Item Information identification model may refer to following embodiment illustrated in fig. 5,
Herein first without being described in detail.
In step s 103: typing is carried out according to the Item Information.
In one embodiment, described image information is identified when based on Item Information identification model trained in advance,
After obtaining the matched Item Information of article to be logged, which can be entered into the cash register system of businessman.
In one embodiment, above-mentioned businessman includes restaurant, fresh food supermarket, retail shop etc., the present embodiment to this without
It limits.Correspondingly, above-mentioned article to be logged may include vegetable, veterinary antibiotics, seafood and various retail items etc., this reality
Example is applied to this without limiting.
Seen from the above description, the present embodiment passes through the image information for obtaining article to be logged, and based on training in advance
Item Information identification model identifies described image information, obtains the matched Item Information of article to be logged, in turn
Typing is carried out according to the Item Information, due to the image information automatic identification Item Information based on article to be logged and is recorded
Enter, without manual mode or voice mode typing, thus the entry time of Item Information can be reduced, improves typing article
The efficiency of information, and the accuracy of typing Item Information can be promoted.
Fig. 2 is the process for how obtaining the matched Item Information of article to be logged shown in one exemplary embodiment of the application
Figure;The present embodiment carries out example for how obtaining the matched Item Information of article to be logged on the basis of the above embodiments
Property explanation.As shown in Fig. 2, being believed based on Item Information identification model trained in advance described image described in above-mentioned steps S102
Breath is identified, is obtained the matched Item Information of article to be logged, be may comprise steps of S201-S202:
In step s 201, by described image information input into Item Information identification model trained in advance, it is based on institute
The output result for stating Item Information identification model determines the probability of the article to be logged Yu each alternative article information matches.
In one embodiment, after obtaining the image information of article to be logged, which can be input in advance
In trained Item Information identification model, the object to be logged is determined with the output result based on the Item Information identification model
The probability of product and each alternative article information matches.
In one embodiment, the input of above-mentioned Item Information identification model is the image information of article to be logged.
In one embodiment, the output result of upper Item Information identification model can be above-mentioned image information and each alternative
Each matched probability of Item Information in Item Information.For example, it after obtaining the image information of article to be logged, can incite somebody to action
The image information of the article to be logged is input in Item Information identification model trained in advance, and then the available image is believed
Breath respectively with alternative article information A, alternative article information B and the matched probability P of alternative article information Ca、PbAnd Pc。
In one embodiment, it is above-mentioned by described image information input into Item Information identification model trained in advance, base
The general of the article to be logged and each alternative article information matches is determined in the output result of the Item Information identification model
The mode of rate is referring also to example shown in following Fig. 5, herein first without being described in detail.
In step S202, the matched Item Information of article to be logged according to the determine the probability.
In one embodiment, when the output result based on the Item Information identification model determine the article to be logged with
After the probability of each alternative article information matches, can according to each probability, determined from each alternative article information it is above-mentioned to
The matched Item Information of typing article.
In one embodiment, above-mentioned each probability can be ranked up according to the size of probability value, and then can be based on
The result of the sequence determines the matched Item Information of article to be logged.
For example, when obtaining the article to be logged and alternative article information A, alternative article information B and alternative article is believed
It ceases C and distinguishes matched probability Pa、PbAnd PcIt afterwards, can be according to the size of probability value to Pa、PbAnd PcIt is ranked up, obtains sequence knot
Fruit such as Pa>Pb>Pc, and then can be according to the ranking results, by alternative article information A, alternative article information B and alternative article
One or more of information C is determined as the matched Item Information of article to be logged.
In one embodiment, the mode of the above-mentioned matched Item Information of article to be logged according to the determine the probability is also
It can be found in following Fig. 3 or embodiment illustrated in fig. 4, herein first without being described in detail.
Seen from the above description, the present embodiment is by identifying described image information input to Item Information trained in advance
In model, the article to be logged and each alternative article information are determined based on the output result of the Item Information identification model
Matched probability, and the matched Item Information of article to be logged according to the determine the probability, i.e., according to the size of probability value
The probability is ranked up, and the matched Item Information of article to be logged is determined based on the result of the sequence, it can be with
Realization accurately determines the matched Item Information of article to be logged, can provide foundation for subsequent automatic input Item Information, can
To realize the entry time for reducing Item Information, the efficiency of typing Item Information is improved, and typing Item Information can be promoted
Accuracy.
Fig. 3 is how based on the result of the sequence to determine the object to be logged shown in one exemplary embodiment of the application
The flow chart of the matched Item Information of product;The present embodiment on the basis of the above embodiments with how the result based on the sequence
It determines and illustrates for the matched Item Information of article to be logged.As shown in figure 3, implementing shown in above-mentioned Fig. 2
Result described in example based on the sequence determines the matched Item Information of article to be logged, can also include the following steps
S301-S302:
In step S301, the maximum alternative article information of probability value is determined based on the result of the sequence.
In step s 302, that the maximum alternative article information of the probability value is determined as the article to be logged is matched
Item Information.
In one embodiment, after obtaining the ranking results of each probability, probability value maximum can be determined according to the result
Alternative article information, and then the maximum alternative article information of the probability value can be determined as above-mentioned article to be logged matching
Item Information.
For example, when obtaining ranking results such as Pa>Pb>PcAfterwards, the maximum alternative article letter of probability value can be determined
Breath, i.e. PaCorresponding alternative article information A, and then it is matched alternative article information A can be determined as to above-mentioned article to be logged
Item Information.
Seen from the above description, the present embodiment is by determining the maximum alternative article of probability value based on the result of the sequence
Information, and the maximum alternative article information of the probability value is determined as the matched Item Information of article to be logged, can be with
It realizes that the result based on the sequence determines the matched Item Information of article to be logged, and then is subsequent automatic input article
Information provides foundation, and the entry time for reducing Item Information may be implemented, improve the efficiency of typing Item Information, and can be promoted
The accuracy of typing Item Information.
Fig. 4 be shown in the application another exemplary embodiment how based on the sequence result determine it is described to be logged
The flow chart of the matched Item Information of article;The present embodiment on the basis of the above embodiments with how the knot based on the sequence
Fruit illustrates for determining the matched Item Information of article to be logged.As shown in figure 4, real shown in above-mentioned Fig. 2
It applies the result described in example based on the sequence and determines the matched Item Information of article to be logged, can also include following step
Rapid S401-S402:
In step S401, the result based on the sequence exports multiple alternative article information and selects for user.
In one embodiment, it after obtaining the ranking results of each probability, can be exported based on the result of the sequence more
A alternative article information is selected for user.
For example, when obtaining ranking results such as Pa>Pb>PcIt afterwards, can be according to probability value to the first two of maximum probability
Alternative article information is exported, for selection by the user.
In one embodiment, when the mode that the first two alternative article information to maximum probability is exported can be by developing
Personnel are configured according to actual needs, such as the two alternative article information and corresponding probability are passed through terminal device exhibition
Show to user, so that user selects suitable Item Information from two alternative article information.
In step S402, the alternative article information that the user selects is determined as the matched object of article to be logged
Product information.
It in one embodiment, can after the result based on the sequence, which exports multiple alternative article information, to be selected for user
It is determined as the matched Item Information of article to be logged to lift the alternative article information that user selects to you.
For example, if by alternative article information A, alternative article information B and corresponding probability Pa、PaShow user
Afterwards, if user selects alternative article information B, alternative article information B can be determined as to the matched object of article to be logged
Product information.
Seen from the above description, the present embodiment by result based on the sequence export multiple alternative article information for
Family selection, and is determined as the matched Item Information of article to be logged for the alternative article information that the user selects, can be with
Realize that the result based on the sequence determines the matched Item Information of article to be logged, it is accurate that raising Item Information determines
Property, and then foundation is provided for subsequent automatic input Item Information, the entry time for reducing Item Information may be implemented, improve typing
The efficiency of Item Information, and the accuracy of typing Item Information can be promoted.
Fig. 5 is the process of the method for how training Item Information identification model shown in one exemplary embodiment of the application
Figure;As shown in figure 5, on the basis of the above embodiments, this method can also include according to following steps S501-S503 training object
Product information identification model:
In step S501, multiple sample item image informations are obtained.
In one embodiment, in order to train Item Information identification model, the image information of available great amount of samples article.
It is worth noting that above-mentioned sample item image information can be the original shot to each sample article
Beginning image information, alternatively, can be the image information obtained after being pre-processed to the original image information of each sample article.
It is worth noting that above-mentioned pretreated mode can be chosen according to actual needs by developer, such as will
Original image processing is the image information etc. with predetermined width, height and pixel format, and the present embodiment is to this without limiting.
In step S502, the corresponding Item Information of each sample item image information is demarcated.
In one embodiment, after obtaining multiple sample item image informations, each sample article figure can be demarcated
As the corresponding Item Information of information.
In one embodiment, the particular content of above-mentioned Item Information can be carried out by developer according to actual business requirement
Setting, be such as set as the title of article, number, classification, attribute, price, in feature at least one of, the present embodiment to this not into
Row limits.
In one embodiment, the mode of the corresponding Item Information of each sample item image information of above-mentioned calibration can be with
It is chosen, is such as demarcated using manual type or unartificial mode, the present embodiment pair according to actual needs by developer
This is without limiting.
In step S503, using the multiple sample item image information and corresponding Item Information as training set, instruction
Practice Item Information identification model.
It in one embodiment, can be by institute after demarcating each sample item image information corresponding Item Information
Multiple sample item image informations and corresponding Item Information are stated as training set, training Item Information identification model.
In one embodiment, the type of above-mentioned Item Information identification model can be carried out according to actual needs by developer
It chooses, such as chooses the deep neural network model based on CNN, the present embodiment is to this without limiting.
In one embodiment, by the multiple sample item image information and corresponding article described in above-mentioned steps S503
Information may comprise steps of S5031-S5032 as training set, training Item Information identification model:
In step S5031, the deep neural network model based on convolutional neural networks CNN is constructed, and is the depth
A loss function is arranged in neural network model, and the deep neural network model has n output, and n is the sum of Item Information
Amount.
In one embodiment, above-mentioned deep neural network model can include but is not limited to three-layer coil product neural network CNN,
The present embodiment is to this without limiting.
It is worth noting that above-mentioned loss function can carry out free setting according to actual business requirement by developer,
It is such as set as intersecting entropy function, the present embodiment is to this without limiting.
It is right using the multiple sample item image information and corresponding Item Information as training set in step S5032
The deep neural network model carries out successive ignition training, until the numerical value of the loss function is less than preset threshold.
It is worth noting that the preset threshold of above-mentioned loss function can be carried out by developer according to actual business requirement
Freely it is arranged, the present embodiment is to this without limiting.
On this basis, described in the output result determination described in above-mentioned steps S201 based on the Item Information identification model
The probability of article to be logged and each alternative article information matches, can also include the following steps S2011-S2013:
In step S2011, the output result of the Item Information identification model is normalized.
In one embodiment, the output result of above-mentioned Item Information identification model can be used to indicate that article and object to be logged
The matching degree parameter of product information.
As an example it is assumed that there is currently n vegetables to classify, be followed successively by Sichuan cuisine, Guangdong dishes, Shandong cuisine ..., North-east China cuisine etc., when
After the image information for obtaining article to be logged (e.g., Sichuan-style pork), which is input to above-mentioned Item Information identification model
In, the output result of available model is for example are as follows: { 1.5,0.2,0.4 ..., 0.3 }, i.e. expression article " Sichuan-style pork " and river
Dish, Guangdong dishes, Shandong cuisine ..., the matching degree parameter of North-east China cuisine be followed successively by 1.5,0.2,0.4 ..., 0.3.It on this basis, can be with
The output result is normalized, and subsequent step is executed based on the result after normalized.
In step S2012, the depth mind is enabled in the result after the normalized is determined based on argmax function
The maximum result of output result through deep neural network in network model.
It is in step S2013, the output result for enabling deep neural network is maximum as a result, being determined as described wait record
Enter the probability of article Yu each alternative article information matches.
Seen from the above description, the present embodiment is by obtaining multiple sample item image informations, and demarcates each sample
The corresponding Item Information of this item image information, and then the multiple sample item image information and corresponding Item Information are made
For training set, training Item Information identification model may be implemented to train the Item Information identification model based on sample image, into
And the matched Item Information of article to be logged can be obtained based on trained Item Information identification model to be subsequent, and according to institute
It states Item Information and typing offer foundation is provided, due to the image information automatic identification Item Information based on article to be logged and carry out
Typing without manual mode or voice mode typing, thus can reduce the entry time of Item Information, improve typing object
The efficiency of product information, and the accuracy of typing Item Information can be promoted.
Fig. 6 is a kind of structural block diagram of the device of typing Item Information shown in one exemplary embodiment of the application;Such as Fig. 6
Shown, which includes: image information acquisition module 110, Item Information acquisition module 120 and Item Information recording module
130, in which:
Image information acquisition module 110, for obtaining the image information of article to be logged;
Item Information obtains module 120, for based on Item Information identification model trained in advance to described image information
It is identified, obtains the matched Item Information of article to be logged;
Item Information recording module 130, for carrying out typing according to the Item Information.
Seen from the above description, the present embodiment passes through the image information for obtaining article to be logged, and based on training in advance
Item Information identification model identifies described image information, obtains the matched Item Information of article to be logged, in turn
Typing is carried out according to the Item Information, due to the image information automatic identification Item Information based on article to be logged and is recorded
Enter, without manual mode or voice mode typing, thus the entry time of Item Information can be reduced, improves typing article
The efficiency of information, and the accuracy of typing Item Information can be promoted.
Fig. 7 is a kind of structural block diagram of the device of typing Item Information shown in the application another exemplary embodiment;Its
In, image information acquisition module 210, Item Information obtain module 220 and Item Information recording module 230 and 6 institute of earlier figures
Show that image information acquisition module 110 in embodiment, Item Information obtain module 120 and Item Information recording module 130
Function is identical, herein without repeating.As shown in fig. 7, Item Information obtains module 220, may include:
Informational probability acquiring unit 221, for described image information input to Item Information trained in advance to be identified mould
In type, the article to be logged and each alternative article information are determined based on the output result of the Item Information identification model
The probability matched;
Item Information obtaining unit 222 is used for the matched Item Information of article to be logged according to the determine the probability.
In one embodiment, Item Information obtaining unit 222 can be also used for:
The probability is ranked up according to the size of probability value;
The matched Item Information of article to be logged is determined based on the result of the sequence.
In one embodiment, Item Information obtaining unit 222 can be also used for:
The maximum alternative article information of probability value is determined based on the result of the sequence;
The maximum alternative article information of the probability value is determined as the matched Item Information of article to be logged.
In one embodiment, Item Information obtaining unit 222 can be also used for:
Result based on the sequence exports multiple alternative article information and selects for user;
The alternative article information that the user selects is determined as the matched Item Information of article to be logged.
In one embodiment, above-mentioned apparatus can also include identification model training module 240;
Identification model training module 240 may include:
Sample image acquiring unit 241, for obtaining multiple sample item image informations;
Item Information demarcates unit 242, for demarcating the corresponding Item Information of each sample item image information;
Identification model training unit 243, for making the multiple sample item image information and corresponding Item Information
For training set, training Item Information identification model.
In one embodiment, identification model training unit 243 can be also used for:
The deep neural network model based on convolutional neural networks CNN is constructed, and is set for the deep neural network model
A loss function is set, the deep neural network model has n output, and n is the total quantity of Item Information.
Using the multiple sample item image information and corresponding Item Information as training set, to the depth nerve net
Network model carries out successive ignition training, until the numerical value of the loss function is less than preset threshold.
On this basis, informational probability acquiring unit 221 can be also used for:
The output result of the Item Information identification model is normalized;
It is enabled in result after determining the normalized based on argmax function deep in the deep neural network model
Spend the maximum result of output result of neural network;
By the output result for enabling deep neural network it is maximum as a result, be determined as the article to be logged with it is each standby
Select the matched probability of Item Information.
It is worth noting that all the above alternatives, can form the optional reality of the disclosure using any combination
Example is applied, this is no longer going to repeat them.
On the other hand, present invention also provides a kind of computer readable storage medium, storage medium is stored with computer journey
Sequence, the method that computer program is used to execute the typing Item Information that above-mentioned FIG. 1 to FIG. 5 illustrated embodiment provides.
For device embodiment, since it corresponds essentially to embodiment of the method, so related place is referring to method reality
Apply the part explanation of example.The apparatus embodiments described above are merely exemplary, wherein described be used as separation unit
The unit of explanation may or may not be physically separated, and component shown as a unit can be or can also be with
It is not physical unit, it can it is in one place, or may be distributed over multiple network units.It can be according to actual
The purpose for needing to select some or all of the modules therein to realize application scheme.Those of ordinary skill in the art are not paying
Out in the case where creative work, it can understand and implement.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to its of the application
Its embodiment.This application is intended to cover any variations, uses, or adaptations of the application, these modifications, purposes or
Person's adaptive change follows the general principle of the application and including the undocumented common knowledge in the art of the application
Or conventional techniques.The description and examples are only to be considered as illustrative, and the true scope and spirit of the application are by following
Claim is pointed out.
It should also be noted that, the terms "include", "comprise" or its any other variant are intended to nonexcludability
It include so that the process, method, commodity or the equipment that include a series of elements not only include those elements, but also to wrap
Include other elements that are not explicitly listed, or further include for this process, method, commodity or equipment intrinsic want
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including described want
There is also other identical elements in the process, method of element, commodity or equipment.
The foregoing is merely the preferred embodiments of the application, not to limit the application, all essences in the application
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the application protection.
Claims (14)
1. a kind of method of typing Item Information characterized by comprising
Obtain the image information of article to be logged;
Described image information is identified based on Item Information identification model trained in advance, obtains the article to be logged
The Item Information matched;
Typing is carried out according to the Item Information.
2. the method according to claim 1, wherein described based on Item Information identification model pair trained in advance
Described image information is identified, the matched Item Information of article to be logged is obtained, comprising:
By described image information input into Item Information identification model trained in advance, it is based on the Item Information identification model
Output result determine the probability of the article to be logged Yu each alternative article information matches;
The matched Item Information of article to be logged according to the determine the probability.
3. according to the method described in claim 2, it is characterized in that, the article to be logged according to the determine the probability
The Item Information matched, comprising:
The probability is ranked up according to the size of probability value;
The matched Item Information of article to be logged is determined based on the result of the sequence.
4. according to the method described in claim 3, it is characterized in that, the result based on the sequence determine it is described to be logged
The matched Item Information of article, comprising:
The maximum alternative article information of probability value is determined based on the result of the sequence;
The maximum alternative article information of the probability value is determined as the matched Item Information of article to be logged.
5. according to the method described in claim 3, it is characterized in that, the result based on the sequence determine it is described to be logged
The matched Item Information of article, comprising:
Result based on the sequence exports multiple alternative article information and selects for user;
The alternative article information that the user selects is determined as the matched Item Information of article to be logged.
6. according to the described in any item methods of claim 2-5, which is characterized in that the method also includes being in advance based on following step
Suddenly the Item Information identification model is trained:
Obtain multiple sample item image informations;
Demarcate the corresponding Item Information of each sample item image information;
Using the multiple sample item image information and corresponding Item Information as training set, training Item Information identifies mould
Type.
7. according to the method described in claim 6, it is characterized in that, described by the multiple sample item image information and correspondence
Item Information as training set, training Item Information identification model, comprising:
The deep neural network model based on convolutional neural networks CNN is constructed, and is deep neural network model setting one
Loss function, the deep neural network model have n output, and n is the total quantity of Item Information;
Using the multiple sample item image information and corresponding Item Information as training set, to the deep neural network mould
Type carries out successive ignition training, until the numerical value of the loss function is less than preset threshold.
8. the method according to the description of claim 7 is characterized in that the output knot based on the Item Information identification model
Fruit determines the probability of the article to be logged Yu each alternative article information matches, comprising:
The output result of the Item Information identification model is normalized;
The output knot of the deep neural network model is enabled in result after determining the normalized based on argmax function
The maximum result of fruit;
Enable the output result maximum as a result, being determined as the article to be logged and each alternative article information matches for described
Probability.
9. a kind of device of typing Item Information characterized by comprising
Image information acquisition module, for obtaining the image information of article to be logged;
Item Information obtains module, for being known based on Item Information identification model trained in advance to described image information
Not, the matched Item Information of article to be logged is obtained;
Item Information recording module, for carrying out typing according to the Item Information.
10. device according to claim 9, which is characterized in that the Item Information obtains module, comprising:
Informational probability acquiring unit, for by described image information input into Item Information identification model trained in advance, base
The general of the article to be logged and each alternative article information matches is determined in the output result of the Item Information identification model
Rate;
Item Information obtaining unit is used for the matched Item Information of article to be logged according to the determine the probability.
11. device according to claim 10, which is characterized in that the Item Information obtaining unit is also used to:
The probability is ranked up according to the size of probability value;
The matched Item Information of article to be logged is determined based on the result of the sequence.
12. device according to claim 11, which is characterized in that the Item Information obtaining unit is also used to:
The maximum alternative article information of probability value is determined based on the result of the sequence;
The maximum alternative article information of the probability value is determined as the matched Item Information of article to be logged.
13. device according to claim 11, which is characterized in that the Item Information obtaining unit is also used to:
Result based on the sequence exports multiple alternative article information and selects for user;
The alternative article information that the user selects is determined as the matched Item Information of article to be logged.
14. a kind of computer readable storage medium, which is characterized in that the storage medium is stored with computer program, the meter
The method that calculation machine program is used to execute any typing Item Information of the claims 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910132029.8A CN109902733A (en) | 2019-02-22 | 2019-02-22 | The method, apparatus and storage medium of typing Item Information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910132029.8A CN109902733A (en) | 2019-02-22 | 2019-02-22 | The method, apparatus and storage medium of typing Item Information |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109902733A true CN109902733A (en) | 2019-06-18 |
Family
ID=66945217
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910132029.8A Pending CN109902733A (en) | 2019-02-22 | 2019-02-22 | The method, apparatus and storage medium of typing Item Information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109902733A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111783574A (en) * | 2020-06-17 | 2020-10-16 | 李利明 | Meal image recognition method and device and storage medium |
CN111907829A (en) * | 2020-06-18 | 2020-11-10 | 戴纳智慧医疗科技有限公司 | Management method for sample transfer and storage box combination |
Citations (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102063616A (en) * | 2010-12-30 | 2011-05-18 | 上海电机学院 | Automatic identification system and method for commodities based on image feature matching |
CN103236007A (en) * | 2013-04-26 | 2013-08-07 | 孙鑫 | On-line poster generation system and on-line poster generation method |
CN105005923A (en) * | 2015-07-07 | 2015-10-28 | 苏州金禾通软件有限公司 | Unique identifier based commodity information processing method and apparatus |
CN105681918A (en) * | 2015-09-16 | 2016-06-15 | 乐视致新电子科技(天津)有限公司 | Method and system for presenting article relevant information in video stream |
CN106980846A (en) * | 2017-05-01 | 2017-07-25 | 刘至键 | A kind of auto parts machinery identifying device based on depth convolutional network |
CN107077659A (en) * | 2016-09-26 | 2017-08-18 | 达闼科技(北京)有限公司 | A kind of intelligent inventory management system, server, method, terminal and program product |
CN107256246A (en) * | 2017-06-06 | 2017-10-17 | 西安工程大学 | PRINTED FABRIC image search method based on convolutional neural networks |
CN107292248A (en) * | 2017-06-05 | 2017-10-24 | 广州诚予国际市场信息研究有限公司 | A kind of merchandise control method and system based on image recognition technology |
CN107368614A (en) * | 2017-09-12 | 2017-11-21 | 重庆猪八戒网络有限公司 | Image search method and device based on deep learning |
CN107525342A (en) * | 2017-07-26 | 2017-12-29 | 合肥美的智能科技有限公司 | Item Information input method, system and refrigerator |
CN107705066A (en) * | 2017-09-15 | 2018-02-16 | 广州唯品会研究院有限公司 | Information input method and electronic equipment during a kind of commodity storage |
CN107729900A (en) * | 2017-09-15 | 2018-02-23 | 广州唯品会研究院有限公司 | It is a kind of that the method and apparatus for completing typing information completion is extracted using picture attribute |
CN107861972A (en) * | 2017-09-15 | 2018-03-30 | 广州唯品会研究院有限公司 | The method and apparatus of the full result of display of commodity after a kind of user's typing merchandise news |
CN107861971A (en) * | 2017-09-15 | 2018-03-30 | 广州唯品会研究院有限公司 | A kind of product search method and device |
CN108256474A (en) * | 2018-01-17 | 2018-07-06 | 百度在线网络技术(北京)有限公司 | For identifying the method and apparatus of vegetable |
CN108596137A (en) * | 2018-05-02 | 2018-09-28 | 济南浪潮高新科技投资发展有限公司 | A kind of commodity scanning record method based on image recognition algorithm |
CN108806133A (en) * | 2018-06-08 | 2018-11-13 | 准动网络科技(厦门)有限公司 | Self-service cash method and system |
CN108830147A (en) * | 2018-05-04 | 2018-11-16 | 广州图匠数据科技有限公司 | A kind of commodity on shelf price recognition methods based on image recognition, device and system |
CN109190444A (en) * | 2018-07-02 | 2019-01-11 | 南京大学 | A kind of implementation method of the lane in which the drivers should pay fees vehicle feature recognition system based on video |
CN109359541A (en) * | 2018-09-17 | 2019-02-19 | 南京邮电大学 | A kind of sketch face identification method based on depth migration study |
-
2019
- 2019-02-22 CN CN201910132029.8A patent/CN109902733A/en active Pending
Patent Citations (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102063616A (en) * | 2010-12-30 | 2011-05-18 | 上海电机学院 | Automatic identification system and method for commodities based on image feature matching |
CN103236007A (en) * | 2013-04-26 | 2013-08-07 | 孙鑫 | On-line poster generation system and on-line poster generation method |
CN105005923A (en) * | 2015-07-07 | 2015-10-28 | 苏州金禾通软件有限公司 | Unique identifier based commodity information processing method and apparatus |
CN105681918A (en) * | 2015-09-16 | 2016-06-15 | 乐视致新电子科技(天津)有限公司 | Method and system for presenting article relevant information in video stream |
CN107077659A (en) * | 2016-09-26 | 2017-08-18 | 达闼科技(北京)有限公司 | A kind of intelligent inventory management system, server, method, terminal and program product |
CN106980846A (en) * | 2017-05-01 | 2017-07-25 | 刘至键 | A kind of auto parts machinery identifying device based on depth convolutional network |
CN107292248A (en) * | 2017-06-05 | 2017-10-24 | 广州诚予国际市场信息研究有限公司 | A kind of merchandise control method and system based on image recognition technology |
CN107256246A (en) * | 2017-06-06 | 2017-10-17 | 西安工程大学 | PRINTED FABRIC image search method based on convolutional neural networks |
CN107525342A (en) * | 2017-07-26 | 2017-12-29 | 合肥美的智能科技有限公司 | Item Information input method, system and refrigerator |
CN107368614A (en) * | 2017-09-12 | 2017-11-21 | 重庆猪八戒网络有限公司 | Image search method and device based on deep learning |
CN107705066A (en) * | 2017-09-15 | 2018-02-16 | 广州唯品会研究院有限公司 | Information input method and electronic equipment during a kind of commodity storage |
CN107729900A (en) * | 2017-09-15 | 2018-02-23 | 广州唯品会研究院有限公司 | It is a kind of that the method and apparatus for completing typing information completion is extracted using picture attribute |
CN107861972A (en) * | 2017-09-15 | 2018-03-30 | 广州唯品会研究院有限公司 | The method and apparatus of the full result of display of commodity after a kind of user's typing merchandise news |
CN107861971A (en) * | 2017-09-15 | 2018-03-30 | 广州唯品会研究院有限公司 | A kind of product search method and device |
CN108256474A (en) * | 2018-01-17 | 2018-07-06 | 百度在线网络技术(北京)有限公司 | For identifying the method and apparatus of vegetable |
CN108596137A (en) * | 2018-05-02 | 2018-09-28 | 济南浪潮高新科技投资发展有限公司 | A kind of commodity scanning record method based on image recognition algorithm |
CN108830147A (en) * | 2018-05-04 | 2018-11-16 | 广州图匠数据科技有限公司 | A kind of commodity on shelf price recognition methods based on image recognition, device and system |
CN108806133A (en) * | 2018-06-08 | 2018-11-13 | 准动网络科技(厦门)有限公司 | Self-service cash method and system |
CN109190444A (en) * | 2018-07-02 | 2019-01-11 | 南京大学 | A kind of implementation method of the lane in which the drivers should pay fees vehicle feature recognition system based on video |
CN109359541A (en) * | 2018-09-17 | 2019-02-19 | 南京邮电大学 | A kind of sketch face identification method based on depth migration study |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111783574A (en) * | 2020-06-17 | 2020-10-16 | 李利明 | Meal image recognition method and device and storage medium |
CN111783574B (en) * | 2020-06-17 | 2024-02-23 | 李利明 | Meal image recognition method, device and storage medium |
CN111907829A (en) * | 2020-06-18 | 2020-11-10 | 戴纳智慧医疗科技有限公司 | Management method for sample transfer and storage box combination |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104598445B (en) | Automatically request-answering system and method | |
CN112733042B (en) | Recommendation information generation method, related device and computer program product | |
CN109831572A (en) | Chat picture control method, device, computer equipment and storage medium | |
CN110728298A (en) | Multi-task classification model training method, multi-task classification method and device | |
CN108345692A (en) | A kind of automatic question-answering method and system | |
CN110245257B (en) | Push information generation method and device | |
CN107301578A (en) | Obtain and recognize the method and its device of customer information | |
CN110163647A (en) | A kind of data processing method and device | |
CN106339507A (en) | Method and device for pushing streaming media message | |
CN108877782A (en) | Audio recognition method and device | |
CN110291538A (en) | Filter the image recognition of image classification output distribution | |
CN108960316A (en) | Method and apparatus for generating model | |
CN109857844A (en) | Intension recognizing method, device, electronic equipment based on dialog text of ordering | |
CN110222728A (en) | The training method of article discrimination model, system and article discrimination method, equipment | |
CN108509499A (en) | A kind of searching method and device, electronic equipment | |
CN108268450B (en) | Method and apparatus for generating information | |
US20160132788A1 (en) | Methods and systems for creating a classifier capable of predicting personality type of users | |
CN108595493A (en) | Method for pushing and device, storage medium, the electronic device of media content | |
CN110147745A (en) | A kind of key frame of video detection method and device | |
CN109388697A (en) | Handle the method, apparatus and electric terminal of question and answer corpus | |
CN109427334A (en) | A kind of man-machine interaction method and system based on artificial intelligence | |
CN107247728B (en) | Text processing method and device and computer storage medium | |
Metwalli et al. | Food image recognition based on densely connected convolutional neural networks | |
CN109582952A (en) | Poem generation method, device, computer equipment and medium | |
CN109299264A (en) | File classification method, device, computer equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for 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: 20190618 |