CN108171275A - For identifying the method and apparatus of flowers - Google Patents
For identifying the method and apparatus of flowers Download PDFInfo
- Publication number
- CN108171275A CN108171275A CN201810044186.9A CN201810044186A CN108171275A CN 108171275 A CN108171275 A CN 108171275A CN 201810044186 A CN201810044186 A CN 201810044186A CN 108171275 A CN108171275 A CN 108171275A
- Authority
- CN
- China
- Prior art keywords
- flowers
- probability
- images
- recognized
- recognition result
- 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 58
- 238000013527 convolutional neural network Methods 0.000 claims description 27
- 230000006870 function Effects 0.000 claims description 21
- 238000004422 calculation algorithm Methods 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 12
- 244000025254 Cannabis sativa Species 0.000 claims description 11
- 235000013399 edible fruits Nutrition 0.000 claims description 9
- 238000010801 machine learning Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 7
- 230000006854 communication Effects 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 240000005001 Paeonia suffruticosa Species 0.000 description 4
- 235000003889 Paeonia suffruticosa Nutrition 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 235000007516 Chrysanthemum Nutrition 0.000 description 2
- 244000284380 Hibiscus rosa sinensis Species 0.000 description 2
- 241000234435 Lilium Species 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000005291 magnetic effect Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 240000005250 Chrysanthemum indicum Species 0.000 description 1
- 244000189548 Chrysanthemum x morifolium Species 0.000 description 1
- WWUVLNPEWUTCEI-KRWDZBQOSA-N Flowerone Chemical compound C1=C(O)C(CC=C(C)C)=CC([C@H]2OC3=C(O)C(O)=CC(O)=C3C(=O)C2)=C1 WWUVLNPEWUTCEI-KRWDZBQOSA-N 0.000 description 1
- 235000005206 Hibiscus Nutrition 0.000 description 1
- 235000007185 Hibiscus lunariifolius Nutrition 0.000 description 1
- 241001495448 Impatiens <genus> Species 0.000 description 1
- 235000010254 Jasminum officinale Nutrition 0.000 description 1
- 240000005385 Jasminum sambac Species 0.000 description 1
- 241000532927 Lagerstroemia Species 0.000 description 1
- 241000736199 Paeonia Species 0.000 description 1
- 235000006484 Paeonia officinalis Nutrition 0.000 description 1
- 244000131316 Panax pseudoginseng Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 235000000664 Rosa chinensis Nutrition 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000009661 flower growth Effects 0.000 description 1
- YVBUWVMMORNHHZ-UHFFFAOYSA-N flowerone Natural products CC(=C)CCc1cc(ccc1O)C1CC(=O)c2c(O)cc(O)c(O)c2O1 YVBUWVMMORNHHZ-UHFFFAOYSA-N 0.000 description 1
- 240000000971 garden vetch Species 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 229930190166 impatien Natural products 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Mathematical Physics (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the present application discloses the method and apparatus for identifying flowers.One specific embodiment of this method includes:Obtain images to be recognized;By images to be recognized input flowers identification model trained in advance, obtain the first recognition result, wherein, first recognition result includes the probability that there are the flowers under the flowers classification in the flowers category set specified in images to be recognized and the probability there is no flowers, and flowers identification model is used to characterize the correspondence between image and the first recognition result;The first recognition result based on gained generates the second recognition result, and exports the second recognition result.The embodiment realizes the identification to flowers.
Description
Technical field
The invention relates to field of computer technology, and in particular to Internet technical field is more particularly, to known
The method and apparatus of other flowers.
Background technology
With being on the increase for flower variety, people are by being visually typically only capable to identify the flowers of a small number of kinds.Therefore,
User is helped to carry out flowers to be identified as a kind of demand.Moreover, flowers identification is also applied to a variety of different applied fields
Scape, such as the check-out flow of florist's shop, flower growth condition monitoring etc..
Invention content
The embodiment of the present application proposes the method and apparatus for identifying flowers.
In a first aspect, the embodiment of the present application provides a kind of method for identifying flowers, this method includes:It obtains and waits to know
Other image;By above-mentioned images to be recognized input flowers identification model trained in advance, the first recognition result is obtained, wherein, it is above-mentioned
First recognition result includes the flowers for having under the flowers classification in the flowers category set specified in above-mentioned images to be recognized
Probability and the probability there is no flowers, above-mentioned flowers identification model are used to characterize the corresponding pass between image and the first recognition result
System;The first recognition result based on gained generates the second recognition result, and exports above-mentioned second recognition result.
In some embodiments, above-mentioned flowers identification model is by being trained to obtain to preset convolutional neural networks
, wherein, above-mentioned convolutional neural networks include convolutional layer, pond layer, full articulamentum and loss layer, in above-mentioned convolutional neural networks
Non- first convolutional layer be connected at least one convolutional layer before the above-mentioned non-first convolutional layer.
In some embodiments, above-mentioned flowers identification model trains to obtain by following training step:It obtains preset
Sample image set and label corresponding with each sample image in above-mentioned sample image set, wherein, above-mentioned sample graph
There is the sample image for showing flowers in closing in image set;Using machine learning method, based on above-mentioned sample image set, above-mentioned sample
The label corresponding to each sample image, preset Classification Loss function and back-propagation algorithm in this image collection is to above-mentioned
Convolutional neural networks are trained, and obtain flowers identification model.
In some embodiments, above-mentioned the first recognition result based on gained generates the second recognition result, including:On determining
State in images to be recognized there is no flowers probability whether be gained the first recognition result in maximum probability;It is if not maximum
Probability, then according to numerical values recited, there are the flowers under the flowers classification in above-mentioned flowers category set from above-mentioned images to be recognized
Probability is chosen in the probability of grass, and the title generation second of the flowers classification corresponding to the probability selected and the probability is identified
As a result.
In some embodiments, above-mentioned according to numerical values recited, there are above-mentioned flowers classification collection from above-mentioned images to be recognized
Probability is chosen in the probability of the flowers under flowers classification in conjunction, including:According to the sequence that numerical value is descending, wait to know to above-mentioned
Probability in other image there are the flowers under the flowers classification in above-mentioned flowers category set is ranked up, and obtains probability sequence;
Preset number probability is chosen since the stem of above-mentioned probability sequence.
In some embodiments, above-mentioned according to numerical values recited, there are above-mentioned flowers classification collection from above-mentioned images to be recognized
Probability is chosen in the probability of the flowers under flowers classification in conjunction, is further included:There are above-mentioned flowers from above-mentioned images to be recognized
The probability not less than probability threshold value is chosen in the probability of the flowers under flowers classification in category set.
In some embodiments, above-mentioned the first recognition result based on gained generates the second recognition result, further includes:If
Maximum probability, then generation are used to indicate the text message that flowers are not present in above-mentioned images to be recognized, by above-mentioned text message and
There is no the probability of flowers in above-mentioned images to be recognized to generate the second recognition result.
In some embodiments, the above method further includes:It is deposited above-mentioned images to be recognized as new sample image
Storage.
Second aspect, the embodiment of the present application provide a kind of device for being used to identify flowers, which includes:It obtains single
Member is configured to obtain images to be recognized;Recognition unit is configured to above-mentioned images to be recognized input flowers trained in advance
Identification model obtains the first recognition result, wherein, above-mentioned first recognition result includes having what is specified in above-mentioned images to be recognized
The probability and the probability there is no flowers of the flowers under flowers classification in flowers category set, above-mentioned flowers identification model are used for
Characterize the correspondence between image and the first recognition result;Output unit is configured to the first recognition result based on gained
The second recognition result is generated, and exports above-mentioned second recognition result.
In some embodiments, above-mentioned flowers identification model is by being trained to obtain to preset convolutional neural networks
, wherein, above-mentioned convolutional neural networks include convolutional layer, pond layer, full articulamentum and loss layer, in above-mentioned convolutional neural networks
Non- first convolutional layer be connected at least one convolutional layer before the above-mentioned non-first convolutional layer.
In some embodiments, above-mentioned flowers identification model trains to obtain by following training step:It obtains preset
Sample image set and label corresponding with each sample image in above-mentioned sample image set, wherein, above-mentioned sample graph
There is the sample image for showing flowers in closing in image set;Using machine learning method, based on above-mentioned sample image set, above-mentioned sample
The label corresponding to each sample image, preset Classification Loss function and back-propagation algorithm in this image collection is to above-mentioned
Convolutional neural networks are trained, and obtain flowers identification model.
In some embodiments, above-mentioned output unit includes:Determination subelement is configured to determine above-mentioned images to be recognized
In there is no flowers probability whether be gained the first recognition result in maximum probability;First generation subelement, configuration are used
In if not maximum probability, then according to numerical values recited, there are the flowers in above-mentioned flowers category set from above-mentioned images to be recognized
Choose probability in the probability of flowers under grass classification, and by the title of the flowers classification corresponding to the probability selected and the probability
Generate the second recognition result.
In some embodiments, above-mentioned first generation subelement is further configured to:According to descending suitable of numerical value
Sequence is ranked up the probability there are the flowers under the flowers classification in above-mentioned flowers category set in above-mentioned images to be recognized,
Obtain probability sequence;Preset number probability is chosen since the stem of above-mentioned probability sequence.
In some embodiments, above-mentioned first generation subelement is further configured to:From above-mentioned images to be recognized
There are the probability chosen in the probability of the flowers under the flowers classification in above-mentioned flowers category set not less than probability threshold value.
In some embodiments, above-mentioned output unit further includes:Second generation subelement, if being configured to most probably
Rate, then generation are used to indicate the text message that flowers are not present in above-mentioned images to be recognized, by above-mentioned text message and above-mentioned treat
Identify that there is no the probability of flowers the second recognition results of generation in image.
In some embodiments, above device further includes:Storage unit is configured to using above-mentioned images to be recognized as new
Sample image stored.
The third aspect, the embodiment of the present application provide a kind of electronic equipment, which includes:One or more processing
Device;Storage device, for storing one or more programs;When said one or multiple programs are by said one or multiple processors
It performs so that said one or multiple processors are realized such as the method for realization method reflection any in first aspect.
Fourth aspect, the embodiment of the present application provide a kind of computer readable storage medium, are stored thereon with computer journey
Sequence is realized when above procedure is executed by processor such as the method for realization method reflection any in first aspect.
Method and apparatus provided by the embodiments of the present application for identifying flowers, by the way that acquired images to be recognized is defeated
Enter flowers identification model trained in advance, to obtain the first recognition result, wherein, it is to be identified which includes this
The probability and the probability there is no flowers that there are the flowers under the flowers classification in the flowers category set specified in image.Then
The first recognition result based on gained generates the second recognition result, and defeated second recognition result.So as to be effectively utilized flowers
Identification model obtains the first recognition result and obtains the second recognition result based on the first recognition result, realizes to flower
The identification of grass.
Description of the drawings
By reading the detailed reflection made to non-limiting example made with reference to the following drawings, the application's is other
Feature, objects and advantages will become more apparent upon:
Fig. 1 is that this application can be applied to exemplary system architecture figures therein;
Fig. 2 is the flow chart for being used to identify one embodiment of the method for flowers according to the application;
Fig. 3 is the schematic diagram for being used to identify an application scenarios of the method for flowers according to the application;
Fig. 4 is the structure diagram for being used to identify one embodiment of the device of flowers according to the application;
Fig. 5 is adapted for the structure diagram of the computer system of the electronic equipment for realizing the embodiment of the present application.
Specific embodiment
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is anti-
The specific embodiment reflected is used only for explaining related invention rather than the restriction to the invention.It also should be noted that in order to
Convenient for reflection, illustrated only in attached drawing and invent relevant part with related.
It should be noted that in the absence of conflict, the feature in embodiment and embodiment in the application can phase
Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
Fig. 1 shows the method for being used to identify flowers that can apply the application or the implementation for identifying the device of flowers
The exemplary system architecture 100 of example.
As shown in Figure 1, system architecture 100 can include terminal device 101,102,103, network 104 and server 105.
Network 104 between terminal device 101,102,103 and server 105 provide communication link medium.Network 104 can be with
Including various connection types, such as wired, wireless communication link or fiber optic cables etc..
User can be interacted with using terminal equipment 101,102,103 by network 104 with server 105, to receive or send out
Send message etc..Various telecommunication customer end applications can be installed, such as web browser should on terminal device 101,102,103
It is applied with, searching class, image identification class application etc..
Terminal device 101,102,103 can be various electronic equipments, including but not limited to smart mobile phone, tablet computer,
Pocket computer on knee and desktop computer etc..
Server 105 can be to provide the server of various services.For example, server 105 can from terminal device 101,
102nd, 103 images to be recognized is obtained, and to the images to be recognized analyze etc. processing, and by handling result (such as generation
Second recognition result) feed back to terminal device.
It should be noted that generally being held for the method that identifies flowers by server 105 of being provided of the embodiment of the present application
Row, correspondingly, the device for identifying flowers is generally positioned in server 105.
It should be pointed out that if images to be recognized is server 105 from locally obtaining, then can in system architecture 100
Not include terminal device 101,102,103.
It should be understood that the number of the terminal device, network and server in Fig. 1 is only schematical.According to realization need
Will, can have any number of terminal device, network and server.
With continued reference to Fig. 2, the flow for being used to identify one embodiment of the method for flowers according to the application is shown
200.This is used for the flow 200 for identifying the method for flowers, includes the following steps:
Step 201, images to be recognized is obtained.
In the present embodiment, for identifying electronic equipment (such as the service shown in FIG. 1 of the method for flowers operation thereon
Device 105) images to be recognized can be obtained from the terminal device (such as terminal device shown in FIG. 1 101,102,103) connected.
Above-mentioned electronic equipment can also receive URL (the Uniform Resource of the images to be recognized of terminal device transmission
Locator, uniform resource locator), images to be recognized is obtained according to the URL.Certainly, above-mentioned electronic equipment can also be from local
Obtain images to be recognized.Wherein, which can show the image of flowers or do not show flowers
Image.
Step 202, by images to be recognized input flowers identification model trained in advance, the first recognition result is obtained.
In the present embodiment, acquired images to be recognized can be inputted flowers trained in advance and known by above-mentioned electronic equipment
Other model, obtains the first recognition result.Wherein, which can include the presence of the flower specified in the images to be recognized
The probability and the probability there is no flowers of the flowers under flowers classification in grass category set.The flowers identification model can be used for
Characterize the correspondence between image and the first recognition result.
It should be noted that the summation of each probability in first recognition result can be equal to 1.In addition, flowers classification
Can be the classification for belonging to the kind in kind according to flowers section to divide, flowers classification can for example include tree peony, Chinese herbaceous peony, chrysanthemum, purple
Common vetch flower, Chinese rose, lily, plum blossom, Jasmine etc..Certainly, flowers classification can also be category or the section belonged to according to flowers section in kind
Come the classification divided.For by the flowers classification marked off is belonged to, flowers classification can for example include lilium, Aloe, Du
Cuckoo category, impatiens, Hibiscus, Chrysanthemum, Lagerstroemia etc..
It should be pointed out that flowers identification model can be technical staff based on to great amount of images and the first recognition result
It counts and pre-establishes, be stored with multiple images and the mapping table of the correspondence of the first recognition result.
In some optional realization methods of the present embodiment, flowers identification model can also be by preset convolution
Neural network (Convolutional Neural Network, CNN) is trained.Wherein, the convolutional neural networks
It can be indiscipline or not train the multilayer convolutional neural networks completed.The convolutional neural networks can for example include convolution
Layer, pond layer, full articulamentum and loss layer.In addition, non-first convolutional layer in the convolutional neural networks can with it is non-positioned at this
At least one convolutional layer before first convolutional layer is connected.For example, the non-first convolutional layer can be with the institute before it
There is convolutional layer to be connected;The non-first convolutional layer can also be connected with the part convolutional layer before it.
It should be noted that above-mentioned flowers identification model can be above-mentioned electronic equipment or remotely lead to above-mentioned electronic equipment
What the server of letter connection was trained by performing following training step:
First, preset sample image set and mark corresponding with each sample image in the sample image set are obtained
Label.Wherein, there may be the sample image for showing flowers in sample image set, there may also be the samples for not showing flowers
This image.For showing the sample image of flowers, the label corresponding to the sample image can serve to indicate that the sample image
The flowers classification that the flowers of display are belonged to.For not showing the sample image of flowers, the label corresponding to the sample image
It can serve to indicate that and flowers are not present in the sample image.In addition, the sample in sample image set and the sample image set
Label corresponding to image can be stored in advance in the training step actuating station (such as above-mentioned electronic equipment or with above-mentioned electronics
The server of equipment telecommunication connection) it is local, naturally it is also possible to it is stored in advance in the server that the actuating station is connected, this
Embodiment does not do any restriction to content in this respect.
Then, using machine learning method, based on each sample image institute in sample image set, sample image set
Corresponding label, preset Classification Loss function and back-propagation algorithm are trained above-mentioned convolutional neural networks, are spent
Grass identification model.Here, in the training process, sample image can be inputted above-mentioned convolutional neural networks by above-mentioned actuating station, be obtained
To the first recognition result corresponding with the sample image, above-mentioned actuating station can determine to be somebody's turn to do using preset Classification Loss function
The difference between label corresponding to first recognition result and the sample image, above-mentioned electronic equipment can be adopted according to the difference
The parameter in above-mentioned convolutional neural networks is adjusted with preset back-propagation algorithm.
It should be noted that above-mentioned Classification Loss function can be various loss function (such as the Hinge for classification
Loss functions or Softmax Loss functions etc.).In the training process, Classification Loss function can constrain the side of convolution kernel modification
Formula and direction, trained target are to make the value of Classification Loss function minimum.Thus, the ginseng of convolutional neural networks obtained after training
The value of number as Classification Loss function parameter corresponding when being minimum value.
In addition, above-mentioned back-propagation algorithm is alternatively referred to as error backpropagation algorithm or Back Propagation Algorithm.Reversely pass
The learning process for broadcasting algorithm is made of forward-propagating process and back-propagation process.In feedforward network, input signal is through input
Layer input, is calculated by hidden layer, is exported by output layer.By output valve compared with mark value, if there is error, by error reversely by
Output layer in this process, can utilize gradient descent algorithm (such as stochastic gradient descent algorithm) right to input Es-region propagations
Neuron weights (such as parameter of convolution kernel etc. in convolutional layer) are adjusted.
Step 203, the first recognition result based on gained generates the second recognition result, and exports the second recognition result.
In the present embodiment, after above-mentioned electronic equipment obtains the first recognition result by performing step 202, above-mentioned electronics is set
Standby first recognition result that can be based on generates the second recognition result, and export second recognition result.It is if for example, acquired
Images to be recognized derives from above-mentioned terminal device, then above-mentioned electronic equipment can export second recognition result to above-mentioned terminal
Equipment.If the images to be recognized is locally obtained from above-mentioned electronic equipment, above-mentioned electronic equipment can tie second identification
Fruit is exported to the display screen of above-mentioned electronic equipment or the specified file stored, naturally it is also possible to output to above-mentioned electronic equipment
The server of telecommunication connection.
Here, above-mentioned electronic equipment can first determine in acquired images to be recognized there is no the probability of flowers whether be
Maximum probability in first recognition result, if not maximum probability, then above-mentioned electronic equipment can will be in the images to be recognized
There are the flowers classifications corresponding to the maximum probability in the probability of the flowers under the flowers classification in above-mentioned flowers category set
Title generates the second recognition result.Assuming that the entitled tree peony of the flowers classification corresponding to the maximum probability, then second knowledge
Other result can include " tree peony ".
Optionally, above-mentioned electronic equipment can also be by the name of the flowers classification corresponding to the maximum probability and the maximum probability
Claim generation the second recognition result.Assuming that the maximum probability is 0.912, flowers classification corresponding to the maximum probability it is entitled male
It is red, then second recognition result can include " tree peony -0.912 ".
In some optional realization methods of the present embodiment, if the probability in the images to be recognized there is no flowers is not
Maximum probability in first recognition result, then above-mentioned electronic equipment can be deposited from the images to be recognized according to numerical values recited
Probability is chosen in the probability of the flowers under flowers classification in above-mentioned flowers category set, and the probability selected is general with this
The title of flowers classification corresponding to rate generates the second recognition result.It should be pointed out that when second recognition result includes two
During a Yi Shang probability, the flowers item name in second recognition result can be according to descending suitable of corresponding probability
Sequence arrangement.
For example, above-mentioned electronic equipment can be according to the descending sequence of numerical value, to there are above-mentioned in the images to be recognized
The probability of the flowers under flowers classification in flowers category set is ranked up, and obtains probability sequence.Above-mentioned electronic equipment can be with
Preset number (such as 3 or 5 etc.) a probability is chosen since the stem of the probability sequence.It should be understood that the preset number is can
With what is be adjusted according to actual needs, the present embodiment does not do any restriction to content in this respect.
For another example above-mentioned electronic equipment can there are the flowers classes in above-mentioned flowers category set from the images to be recognized
The probability not less than probability threshold value (such as 0.5 etc.) is chosen in the probability of flowers under not.It should be understood that the probability threshold value is can
With what is be adjusted according to actual needs, the present embodiment does not do any restriction to content in this respect.
In some optional realization methods of the present embodiment, if in the images to be recognized there are the probability of flowers be gained
The first recognition result in maximum probability, then above-mentioned electronic equipment can generate to be used to indicate in the images to be recognized and be not present
The text message of flowers, by there is no the probability of flowers the second recognition results of generation in text information and the images to be recognized.
For example, text information can include " non-flowers ".Assuming that the probability is 0.998, which can include " non-flower
Grass -0.998 ".
In some optional realization methods of the present embodiment, above-mentioned electronic equipment can be using the images to be recognized as new
Sample image stored.In this way, can continuous enlarged sample image quantity, and the new sample image can be used
In the follow-up training flow of above-mentioned flowers identification model, above-mentioned flowers identification model can be made to pass through iteration and more newly arrive raising in advance
Survey accuracy.
With continued reference to Fig. 3, Fig. 3 is the signal for being used to identify the application scenarios of the method for flowers according to the present embodiment
Figure.In the application scenarios of Fig. 3, first, user can show flowers by the terminal device held to server upload
Images to be recognized 301;Then, images to be recognized 301 can be inputted flowers identification model by server, obtain the first identification knot
Fruit;Then, server can be by preceding 5 probability of numerical value maximum and the name of corresponding flowers classification in the first recognition result
Claim generation the second recognition result 302, and the second recognition result 302 is exported to terminal device.Wherein, can be on terminal device
301 and second recognition result 302 of existing images to be recognized.
The method that above-described embodiment of the application provides, is effectively utilized flowers identification model to obtain the first identification knot
Fruit and the second recognition result is obtained based on the first recognition result, realize the identification to flowers.
With further reference to Fig. 4, as the realization to method shown in above-mentioned each figure, this application provides one kind for identifying flower
One embodiment of the device of grass, the device embodiment is corresponding with embodiment of the method shown in Fig. 2, which can specifically answer
For in various electronic equipments.
As shown in figure 4, it is used to identify that the device 400 of flowers to include shown in the present embodiment:Acquiring unit 401, identification are single
Member 402 and output unit 403.Wherein, acquiring unit 401 is configured to obtain images to be recognized;Recognition unit 402 is configured to
By above-mentioned images to be recognized input flowers identification model trained in advance, the first recognition result is obtained, wherein, above-mentioned first identification
As a result include the probability that there are the flowers under the flowers classification in the flowers category set specified in above-mentioned images to be recognized and not
There are the probability of flowers, above-mentioned flowers identification model is used to characterize the correspondence between image and the first recognition result;Output
Unit 403 is configured to the first recognition result based on gained and generates the second recognition result, and exports above-mentioned second recognition result.
In the present embodiment, for identifying in the device 400 of flowers:Acquiring unit 401, recognition unit 402 and output are single
The specific processing of member 403 and its caused technique effect can be respectively with reference to step 201, the steps 202 in 2 corresponding embodiment of figure
With the related description of step 203, details are not described herein.
In some optional realization methods of the present embodiment, above-mentioned flowers identification model can be by preset volume
Product neural network is trained, wherein, the convolutional neural networks can include convolutional layer, pond layer, full articulamentum and
Loss layer, the non-first convolutional layer in the convolutional neural networks and at least one convolutional layer before the non-first convolutional layer
It is connected.
In some optional realization methods of the present embodiment, above-mentioned flowers identification model can be walked by following training
What rapid training obtained:Obtain preset sample image set and corresponding with each sample image in above-mentioned sample image set
Label, wherein, there may be the sample image for showing flowers in above-mentioned sample image set;Utilize machine learning method, base
The label corresponding to each sample image, preset classification damage in above-mentioned sample image set, above-mentioned sample image set
It loses function and back-propagation algorithm is trained above-mentioned convolutional neural networks, obtain flowers identification model.
In some optional realization methods of the present embodiment, above-mentioned output unit 403 can include:Determination subelement
(not shown), be configured to determine above-mentioned images to be recognized in there is no flowers probability whether be gained first identification
As a result the maximum probability in;First generation subelement (not shown), is configured to if not maximum probability, then according to numerical value
Size, it is general there are being chosen in the probability of the flowers under the flowers classification in above-mentioned flowers category set from above-mentioned images to be recognized
Rate, and the title of the flowers classification corresponding to the probability selected and the probability is generated into the second recognition result.
In some optional realization methods of the present embodiment, use can be further configured in above-mentioned first generation subelement
In:According to the sequence that numerical value is descending, to there are the flowers classifications in above-mentioned flowers category set in above-mentioned images to be recognized
Under the probability of flowers be ranked up, obtain probability sequence;It is general that preset number is chosen since the stem of above-mentioned probability sequence
Rate.
In some optional realization methods of the present embodiment, use can also be further configured in above-mentioned first generation subelement
In:It is not small there are being chosen in the probability of the flowers under the flowers classification in above-mentioned flowers category set from above-mentioned images to be recognized
In the probability of probability threshold value.
In some optional realization methods of the present embodiment, above-mentioned output unit 403 can also include:Second generation
Unit (not shown), if being configured to maximum probability, then generation, which is used to indicate in above-mentioned images to be recognized, is not present flower
The text message of grass ties the second identification of probability generation that flowers are not present in above-mentioned text message and above-mentioned images to be recognized
Fruit.
In some optional realization methods of the present embodiment, above device 400 can also include:Storage unit is (in figure
It is not shown), it is configured to store above-mentioned images to be recognized as new sample image.
The device that above-described embodiment of the application provides, is effectively utilized flowers identification model to obtain the first identification knot
Fruit and the second recognition result is obtained based on the first recognition result, realize the identification to flowers.
Below with reference to Fig. 5, it illustrates suitable for being used for realizing the computer system 500 of the electronic equipment of the embodiment of the present application
Structure diagram.Electronic equipment shown in Fig. 5 is only an example, to the function of the embodiment of the present application and should not use model
Shroud carrys out any restrictions.
As shown in figure 5, computer system 500 includes central processing unit (CPU) 501, it can be read-only according to being stored in
Program in memory (ROM) 502 or be loaded into program in random access storage device (RAM) 503 from storage section 508 and
Perform various appropriate actions and processing.In RAM 503, also it is stored with system 500 and operates required various programs and data.
CPU 501, ROM 502 and RAM 503 are connected with each other by bus 504.Input/output (I/O) interface 505 is also connected to always
Line 504.
I/O interfaces 505 are connected to lower component:Importation 506 including keyboard, mouse etc.;It is penetrated including such as cathode
The output par, c 507 of spool (CRT), liquid crystal display (LCD) etc. and loud speaker etc.;Storage section 508 including hard disk etc.;
And the communications portion 509 of the network interface card including LAN card, modem etc..Communications portion 509 via such as because
The network of spy's net performs communication process.Driver 510 is also according to needing to be connected to I/O interfaces 505.Detachable media 511, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on driver 510, as needed in order to be read from thereon
Computer program be mounted into storage section 508 as needed.
Particularly, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart reflection
Software program.For example, embodiment of the disclosure includes a kind of computer program product, including being carried on computer-readable medium
On computer program, which includes for the program code of the method shown in execution flow chart.In such reality
It applies in example, which can be downloaded and installed from network by communications portion 509 and/or from detachable media
511 are mounted.When the computer program is performed by central processing unit (CPU) 501, perform what is limited in the system of the application
Above-mentioned function.
It should be noted that the computer-readable medium shown in the application can be computer-readable signal media or meter
Calculation machine readable storage medium storing program for executing either the two arbitrarily combines.Computer readable storage medium for example can be --- but not
It is limited to --- electricity, magnetic, optical, electromagnetic, system, device or the device of infrared ray or semiconductor or arbitrary above combination.Meter
The more specific example of calculation machine readable storage medium storing program for executing can include but is not limited to:Electrical connection with one or more conducting wires, just
It takes formula computer disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type and may be programmed read-only storage
Device (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device,
Or above-mentioned any appropriate combination.In this application, computer readable storage medium can any include or store journey
The tangible medium of sequence, the program can be commanded the either device use or in connection of execution system, device.And at this
In application, computer-readable signal media can include in a base band or as a carrier wave part propagation data-signal,
Wherein carry computer-readable program code.Diversified forms may be used in the data-signal of this propagation, including but it is unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used for
By instruction execution system, device either device use or program in connection.It is included on computer-readable medium
Program code can be transmitted with any appropriate medium, including but not limited to:Wirelessly, electric wire, optical cable, RF etc. or above-mentioned
Any appropriate combination.
Flow chart and block diagram in attached drawing, it is illustrated that according to the system of the various embodiments of the application, method and computer journey
Architectural framework in the cards, function and the operation of sequence product.In this regard, each box in flow chart or block diagram can generation
The part of one module of table, program segment or code, a part for above-mentioned module, program segment or code include one or more
The executable instruction of logic function as defined in being used to implement.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also be occurred with being different from the sequence marked in attached drawing.For example, two boxes succeedingly represented are practical
On can perform substantially in parallel, they can also be performed in the opposite order sometimes, this is depended on the functions involved.Also
It is noted that the combination of each box in block diagram or flow chart and the box in block diagram or flow chart, can use and perform rule
The group of specialized hardware and computer instruction is realized or can be used to the dedicated hardware based system of fixed functions or operations
It closes to realize.
Being reflected in unit involved in the embodiment of the present application can be realized by way of software, can also be by hard
The mode of part is realized.The unit reflected can also be set in the processor, for example, can be reflected as:A kind of processor packet
Include acquiring unit, recognition unit and output unit.Wherein, the title of these units is not formed under certain conditions to the unit
The restriction of itself, for example, acquiring unit can also be reflected as " unit for obtaining images to be recognized ".
As on the other hand, present invention also provides a kind of computer-readable medium, which can be
Included in the electronic equipment reflected in above-described embodiment;Can also be individualism, and without be incorporated the electronic equipment in.
Above computer readable medium carries one or more program, and when said one or multiple programs, by one, the electronics is set
During standby execution so that the electronic equipment includes:Obtain images to be recognized;Images to be recognized input flowers trained in advance are known
Other model, obtains the first recognition result, wherein, which can include the presence of the flower specified in the images to be recognized
The probability and the probability there is no flowers of the flowers under flowers classification in grass category set, the flowers identification model can be used for
Characterize the correspondence between image and the first recognition result;The first recognition result based on gained generates the second recognition result,
And export second recognition result.
The preferred embodiment and the explanation to institute's application technology principle that above reflection is only the application.People in the art
Member should be appreciated that invention scope involved in the application, however it is not limited to the technology that the specific combination of above-mentioned technical characteristic forms
Scheme, while should also cover in the case where not departing from foregoing invention design, it is carried out by above-mentioned technical characteristic or its equivalent feature
The other technical solutions for arbitrarily combining and being formed.Such as features described above has similar work(with (but not limited to) disclosed herein
The technical solution that the technical characteristic of energy is replaced mutually and formed.
Claims (18)
1. a kind of method for identifying flowers, including:
Obtain images to be recognized;
By images to be recognized input flowers identification model trained in advance, the first recognition result is obtained, wherein, described first
Recognition result includes the probability that there are the flowers under the flowers classification in the flowers category set specified in the images to be recognized
With there is no the probability of flowers, the flowers identification model is used to characterize the correspondence between image and the first recognition result;
The first recognition result based on gained generates the second recognition result, and exports second recognition result.
2. according to the method described in claim 1, wherein, the flowers identification model is by preset convolutional neural networks
It is trained, wherein, the convolutional neural networks include convolutional layer, pond layer, full articulamentum and loss layer, the volume
Non- first convolutional layer in product neural network is connected at least one convolutional layer before the non-first convolutional layer.
3. according to the method described in claim 2, wherein, the flowers identification model is to train to obtain by following training step
's:
Preset sample image set and label corresponding with each sample image in the sample image set are obtained,
In, there is the sample image for showing flowers in the sample image set;
Using machine learning method, based on each sample image institute in the sample image set, the sample image set
Corresponding label, preset Classification Loss function and back-propagation algorithm are trained the convolutional neural networks, are spent
Grass identification model.
4. according to the method described in claim 1, wherein, second identification of the first recognition result generation based on gained is tied
Fruit, including:
Determine in the images to be recognized there is no flowers probability whether be gained the first recognition result in maximum probability;
If not maximum probability, then according to numerical values recited, there are in the flowers category set from the images to be recognized
Choose probability in the probability of flowers under flowers classification, and by the name of the flowers classification corresponding to the probability selected and the probability
Claim generation the second recognition result.
5. according to the method described in claim 4, wherein, described according to numerical values recited, there are institutes from the images to be recognized
It states in the probability of the flowers under the flowers classification in flowers category set and chooses probability, including:
According to the sequence that numerical value is descending, to there are the flowers classifications in the flowers category set in the images to be recognized
Under the probability of flowers be ranked up, obtain probability sequence;
Preset number probability is chosen since the stem of the probability sequence.
6. according to the method described in claim 4, wherein, described according to numerical values recited, there are institutes from the images to be recognized
It states in the probability of the flowers under the flowers classification in flowers category set and chooses probability, further include:
There are chosen not in the probability of the flowers under the flowers classification in the flowers category set from the images to be recognized
Less than the probability of probability threshold value.
7. according to the method described in claim 4, wherein, second identification of the first recognition result generation based on gained is tied
Fruit further includes:
If maximum probability, then generation is used to indicate the text message that flowers are not present in the images to be recognized, by the text
There is no the probability of flowers in this information and the images to be recognized to generate the second recognition result.
8. according to the method described in claim 1, wherein, the method further includes:
It is stored the images to be recognized as new sample image.
9. it is a kind of for identifying the device of flowers, including:
Acquiring unit is configured to obtain images to be recognized;
Recognition unit is configured to, by images to be recognized input flowers identification model trained in advance, obtain the first identification
As a result, wherein, first recognition result includes the flowers for having in the flowers category set specified in the images to be recognized
The probability of flowers under classification and the probability there is no flowers, the flowers identification model are tied for characterizing image and the first identification
Correspondence between fruit;
Output unit is configured to the first recognition result based on gained and generates the second recognition result, and exports described second and know
Other result.
10. device according to claim 9, wherein, the flowers identification model is by preset convolutional Neural net
What network was trained, wherein, the convolutional neural networks include convolutional layer, pond layer, full articulamentum and loss layer, described
Non- first convolutional layer in convolutional neural networks is connected at least one convolutional layer before the non-first convolutional layer.
11. device according to claim 10, wherein, the flowers identification model is trained by following training step
It arrives:
Preset sample image set and label corresponding with each sample image in the sample image set are obtained,
In, there is the sample image for showing flowers in the sample image set;
Using machine learning method, based on each sample image institute in the sample image set, the sample image set
Corresponding label, preset Classification Loss function and back-propagation algorithm are trained the convolutional neural networks, are spent
Grass identification model.
12. device according to claim 9, wherein, the output unit includes:
Determination subelement, be configured to determine the images to be recognized in there is no flowers probability whether be gained first know
Maximum probability in other result;
First generation subelement, is configured to if not maximum probability, then according to numerical values recited, deposit from the images to be recognized
Probability is chosen in the probability of the flowers under flowers classification in the flowers category set, and the probability selected is general with this
The title of flowers classification corresponding to rate generates the second recognition result.
13. device according to claim 12, wherein, the first generation subelement is further configured to:
According to the sequence that numerical value is descending, to there are the flowers classifications in the flowers category set in the images to be recognized
Under the probability of flowers be ranked up, obtain probability sequence;
Preset number probability is chosen since the stem of the probability sequence.
14. device according to claim 12, wherein, the first generation subelement is further configured to:
There are chosen not in the probability of the flowers under the flowers classification in the flowers category set from the images to be recognized
Less than the probability of probability threshold value.
15. device according to claim 12, wherein, the output unit further includes:
Second generation subelement, if being configured to maximum probability, then generation, which is used to indicate in the images to be recognized, is not present
The text message of flowers ties the second identification of probability generation that flowers are not present in the text message and the images to be recognized
Fruit.
16. device according to claim 9, wherein, described device further includes:
Storage unit is configured to store the images to be recognized as new sample image.
17. a kind of electronic equipment, including:
One or more processors;
Storage device, for storing one or more programs,
When one or more of programs are performed by one or more of processors so that one or more of processors are real
Now such as method according to any one of claims 1-8.
18. a kind of computer readable storage medium, is stored thereon with computer program, wherein, described program is executed by processor
Shi Shixian methods for example according to any one of claims 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810044186.9A CN108171275A (en) | 2018-01-17 | 2018-01-17 | For identifying the method and apparatus of flowers |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810044186.9A CN108171275A (en) | 2018-01-17 | 2018-01-17 | For identifying the method and apparatus of flowers |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108171275A true CN108171275A (en) | 2018-06-15 |
Family
ID=62514523
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810044186.9A Pending CN108171275A (en) | 2018-01-17 | 2018-01-17 | For identifying the method and apparatus of flowers |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108171275A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110276405A (en) * | 2019-06-26 | 2019-09-24 | 北京百度网讯科技有限公司 | Method and apparatus for output information |
CN110728300A (en) * | 2019-09-09 | 2020-01-24 | 交控科技股份有限公司 | Method and system for identifying fault type based on turnout action current curve |
CN111709480A (en) * | 2020-06-17 | 2020-09-25 | 北京百度网讯科技有限公司 | Method and device for identifying image category |
CN111814813A (en) * | 2019-04-10 | 2020-10-23 | 北京市商汤科技开发有限公司 | Neural network training and image classification method and device |
CN113569593A (en) * | 2020-04-28 | 2021-10-29 | 京东方科技集团股份有限公司 | Intelligent vase system, flower identification and display method and electronic equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104361348A (en) * | 2014-10-27 | 2015-02-18 | 华南理工大学 | Flower and plant recognition method on intelligent terminal |
CN106960219A (en) * | 2017-03-10 | 2017-07-18 | 百度在线网络技术(北京)有限公司 | Image identification method and device, computer equipment and computer-readable medium |
CN107153844A (en) * | 2017-05-12 | 2017-09-12 | 上海斐讯数据通信技术有限公司 | The accessory system being improved to flowers identifying system and the method being improved |
CN107516128A (en) * | 2017-06-12 | 2017-12-26 | 南京邮电大学 | A kind of flowers recognition methods of the convolutional neural networks based on ReLU activation primitives |
CN107527065A (en) * | 2017-07-25 | 2017-12-29 | 北京联合大学 | A kind of flower variety identification model method for building up based on convolutional neural networks |
-
2018
- 2018-01-17 CN CN201810044186.9A patent/CN108171275A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104361348A (en) * | 2014-10-27 | 2015-02-18 | 华南理工大学 | Flower and plant recognition method on intelligent terminal |
CN106960219A (en) * | 2017-03-10 | 2017-07-18 | 百度在线网络技术(北京)有限公司 | Image identification method and device, computer equipment and computer-readable medium |
CN107153844A (en) * | 2017-05-12 | 2017-09-12 | 上海斐讯数据通信技术有限公司 | The accessory system being improved to flowers identifying system and the method being improved |
CN107516128A (en) * | 2017-06-12 | 2017-12-26 | 南京邮电大学 | A kind of flowers recognition methods of the convolutional neural networks based on ReLU activation primitives |
CN107527065A (en) * | 2017-07-25 | 2017-12-29 | 北京联合大学 | A kind of flower variety identification model method for building up based on convolutional neural networks |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111814813A (en) * | 2019-04-10 | 2020-10-23 | 北京市商汤科技开发有限公司 | Neural network training and image classification method and device |
CN111814813B (en) * | 2019-04-10 | 2024-07-19 | 北京市商汤科技开发有限公司 | Neural network training and image classification method and device |
CN110276405A (en) * | 2019-06-26 | 2019-09-24 | 北京百度网讯科技有限公司 | Method and apparatus for output information |
CN110276405B (en) * | 2019-06-26 | 2022-03-01 | 北京百度网讯科技有限公司 | Method and apparatus for outputting information |
CN110728300A (en) * | 2019-09-09 | 2020-01-24 | 交控科技股份有限公司 | Method and system for identifying fault type based on turnout action current curve |
CN113569593A (en) * | 2020-04-28 | 2021-10-29 | 京东方科技集团股份有限公司 | Intelligent vase system, flower identification and display method and electronic equipment |
CN111709480A (en) * | 2020-06-17 | 2020-09-25 | 北京百度网讯科技有限公司 | Method and device for identifying image category |
CN111709480B (en) * | 2020-06-17 | 2023-06-23 | 北京百度网讯科技有限公司 | Method and device for identifying image category |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108256476A (en) | For identifying the method and apparatus of fruits and vegetables | |
CN108171275A (en) | For identifying the method and apparatus of flowers | |
CN108256474A (en) | For identifying the method and apparatus of vegetable | |
CN107832468B (en) | Demand recognition methods and device | |
CN107168952A (en) | Information generating method and device based on artificial intelligence | |
CN109325541A (en) | Method and apparatus for training pattern | |
CN108090162A (en) | Information-pushing method and device based on artificial intelligence | |
CN109460513A (en) | Method and apparatus for generating clicking rate prediction model | |
CN108038469A (en) | Method and apparatus for detecting human body | |
CN109903112A (en) | Information output method and device | |
CN108171276A (en) | For generating the method and apparatus of information | |
CN108388674A (en) | Method and apparatus for pushed information | |
CN108171203A (en) | For identifying the method and apparatus of vehicle | |
CN108229485A (en) | For testing the method and apparatus of user interface | |
CN110263938A (en) | Method and apparatus for generating information | |
CN109976997A (en) | Test method and device | |
CN108388563A (en) | Information output method and device | |
CN108776692A (en) | Method and apparatus for handling information | |
CN108256591A (en) | For the method and apparatus of output information | |
CN109871311A (en) | A kind of method and apparatus for recommending test case | |
CN109976995A (en) | Method and apparatus for test | |
CN110457476A (en) | Method and apparatus for generating disaggregated model | |
CN108960110A (en) | Method and apparatus for generating information | |
CN110019742A (en) | Method and apparatus for handling information | |
CN109634833A (en) | A kind of Software Defects Predict Methods and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
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: 20180615 |