CN112329116A - Distortion zero space planning design generation method and system based on generation of countermeasure network - Google Patents
Distortion zero space planning design generation method and system based on generation of countermeasure network Download PDFInfo
- Publication number
- CN112329116A CN112329116A CN202011321083.6A CN202011321083A CN112329116A CN 112329116 A CN112329116 A CN 112329116A CN 202011321083 A CN202011321083 A CN 202011321083A CN 112329116 A CN112329116 A CN 112329116A
- Authority
- CN
- China
- Prior art keywords
- network
- generation
- design scheme
- generating
- data set
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Architecture (AREA)
- Civil Engineering (AREA)
- Structural Engineering (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a distortion zero space planning design generation method and a distortion zero space planning design generation system based on generation of a countermeasure network, wherein the method comprises the following steps: establishing a data set of a real distortion null space planning design scheme, and labeling an expected label for each design scheme in the data set; generating a confrontation network structure generation network and a discrimination network based on the deep convolution; generating a true and false judgment result of the first design scheme based on the generation network and the judgment network according to the data set and the expected label; generating a true and false judgment result of the random design scheme based on the discrimination network according to the data set and the expected label; alternately and iteratively training a discrimination network and a generation network; and receiving the expected label form submitted by the user, and generating a second design scheme corresponding to the expected label form by using the trained generation network.
Description
Technical Field
The invention relates to the technical field of building generation design, in particular to a distortion zero space planning design generation method and system based on generation of a countermeasure network.
Background
Along with the acceleration of the urbanization process and the improvement of the economic development level in recent years, the urban land scale is continuously enlarged, and the problem of land resource scarcity is increasingly highlighted. The abnormal space is that in the process of urban building development, some lands have unreasonable phenomena during planning and using. Therefore, some inevitable space waste is inevitably generated between buildings or between communities due to different demands. The inefficient use of these spaces may be due to the following reasons: (1) irregular plots formed between buildings are narrow; (2) due to building construction, the land is polluted; (3) terrain restrictions, such as steep slopes; (4) the urban construction needs to build an elevated structure, particularly more elevated structures in some big cities, and the elevated structures divide roads into different equal parts; (5) the requirements of the original facilities. Resulting in a change in usage.
The rapid development of modern buildings causes the difference of new and old buildings in cities, and the increasing of population number greatly reduces the living space of urban residents. In addition, the traffic land occupies about 10% of the whole urban area, the cost required in the process of repairing the overhead is often too high, but after the overhead is repaired, a lot of land is wasted, and the loss rate of the land is extremely high. The utilization of the distortion null space can break the gap between different buildings, and better ensure the continuity of the city space, thereby creating a more meaningful public space.
With the development of artificial intelligence technology, some researchers have begun to study how artificial intelligence technology can be applied to the field of building design. However, for the distortion space, the requirement is relatively small on one hand, and the space structure is irregular on the other hand, so that no scholars are engaged in the research of the related direction. However, the planning and designing of the distortion null space in a manual mode is time-consuming and labor-consuming, and meanwhile, the profit is small, and a suitable designer is difficult to find. Therefore, the artificial intelligence technology is introduced into the distortion zero space planning design, the labor cost is saved, and the method has important significance.
Disclosure of Invention
The invention provides a distortion null space planning design generation method and system based on generation of a countermeasure network, aiming at the technical problem of utilizing the distortion null space.
In a first aspect, an embodiment of the present application provides a distortion null space planning design generation method based on generation of an antagonistic network, including:
a data set generation step: establishing a data set of a real distortion null space planning design scheme, and labeling an expected label for each design scheme in the data set;
a network construction step: generating a confrontation network structure generation network and a discrimination network based on the deep convolution;
a first result generation step: generating a true and false judgment result of a first design scheme based on the generation network and the discrimination network according to the data set and the expected label;
a random result generation step: generating a true and false judgment result of a random design scheme based on the discrimination network according to the data set and the expected label;
network training: alternately and iteratively training the discrimination network and the generation network;
a second scheme generation step: and receiving a desired label form submitted by a user, and generating a second design scheme corresponding to the desired label form by using the trained generation network.
The method for generating the distorted null space programming design comprises the steps that the generation network is a deep deconvolution neural network, and the discrimination network is a deep convolution neural network.
The method for generating a distorted null-space planning design includes:
a first scheme generation step: inputting the expected label into the generation network, and outputting a corresponding first design scheme;
a first judgment result generation step: and simultaneously inputting the real design scheme, the first design scheme and the expected label corresponding to the first design scheme in the data set into the discrimination network, and outputting a true and false judgment result of the first design scheme.
The method for generating the distorted null space planning design comprises the following steps:
and simultaneously inputting a certain expected label in the data set, a real design scheme corresponding to the expected label and a random real design scheme into the discrimination network, and outputting a true and false judgment result of the random design scheme.
The distortion zero space planning design generation method comprises the following network training steps:
function setting step: setting a loss function of the discrimination network and the generation network during training;
and (3) updating parameters: and updating the parameters of the discrimination network and the generation network.
In a second aspect, an embodiment of the present application provides a distortion-free space planning design generation system based on generation of an antagonistic network, including:
a dataset generation module: establishing a data set of a real distortion null space planning design scheme, and labeling an expected label for each design scheme in the data set;
a network construction module: generating a confrontation network structure generation network and a discrimination network based on the deep convolution;
a first result generation module: generating a true and false judgment result of a first design scheme based on the generation network and the discrimination network according to the data set and the expected label;
a random result generation module: generating a true and false judgment result of a random design scheme based on the discrimination network according to the data set and the expected label;
a network training module: alternately and iteratively training the discrimination network and the generation network;
a second scenario generation module: and receiving a desired label form submitted by a user, and generating a second design scheme corresponding to the desired label form by using the trained generation network.
The above-mentioned distortion null space planning design generation system, wherein, the generation network is a deep deconvolution neural network, and the discrimination network is a deep convolution neural network.
The above distorted null-space planning design generating system, wherein the first result generating module comprises:
a first scheme generation unit: inputting the expected label into the generation network, and outputting a corresponding first design scheme;
a first determination result generation unit: and simultaneously inputting the real design scheme, the first design scheme and the expected label corresponding to the first design scheme in the data set into the discrimination network, and outputting a true and false judgment result of the first design scheme.
The above distorted null space planning design generating system, wherein the random result generating module comprises:
and simultaneously inputting a certain expected label in the data set, a real design scheme corresponding to the expected label and a random real design scheme into the discrimination network, and outputting a true and false judgment result of the random design scheme.
The above distortion null space planning design generating system, wherein the network training module comprises:
a function setting unit: setting a loss function of the discrimination network and the generation network during training;
a parameter updating unit: and updating the parameters of the discrimination network and the generation network.
Compared with the prior art, the invention has the advantages and positive effects that:
1. the scheme effectively utilizes artificial intelligence technology, adopts a deep convolutional neural network framework, plans and designs the teratocardiology space based on the mode of generating the countermeasure network, and generates a proper design scheme according to an expected form of the user teratocardiology space planning. Because spatial structure is irregular, introduce the distortion space planning design with artificial intelligence technique, to practicing thrift the human cost and have the significance, can save designer's human cost on the one hand, on the other hand can save the demand side again and seek designer's time cost, convenient, swift demand side's demand simultaneously.
2. The distortion null space can integrate social space in a certain sense and can also promote the economic development of cities. A large amount of surplus spaces can lead to city view too chaotic, only solve it, just can reduce the destruction to the environment, increase the comfort level in city, improve the economic level in city. The method and the system are utilized to plan and design the distortion null space, so that the planning strength of the city can be improved powerfully, and the city appearance of the city is ensured to be more compact and neat.
Drawings
FIG. 1 is a schematic step diagram of a distortion null space planning design generation method based on generation of a countermeasure network according to the present invention;
FIG. 2 is a flowchart based on step S3 in FIG. 1 according to the present invention;
FIG. 3 is a flowchart based on step S5 in FIG. 1 according to the present invention;
fig. 4 is a block diagram of a distortion null space planning design generation system based on generation of a countermeasure network according to the present invention.
Wherein the reference numerals are:
11. a data set generation module; 12. a network construction module; 13. a first result generation module; 131. a first scheme generation unit; 132. a first judgment result generation unit; 14. a random result generation module; 15. a network training module; 151. a function setting unit; 152. a parameter updating unit; 16. and a second scheme generation module.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application.
It is obvious that the drawings in the following description are only examples or embodiments of the present application, and that it is also possible for a person skilled in the art to apply the present application to other similar contexts on the basis of these drawings without inventive effort. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.
Unless defined otherwise, technical or scientific terms referred to herein shall have the ordinary meaning as understood by those of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The present application is directed to the use of the terms "including," "comprising," "having," and any variations thereof, which are intended to cover non-exclusive inclusions; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Reference to "connected," "coupled," and the like in this application is not intended to be limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The term "plurality" as referred to herein means two or more. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. Reference herein to the terms "first," "second," "third," and the like, are merely to distinguish similar objects and do not denote a particular ordering for the objects.
The present invention is described in detail with reference to the embodiments shown in the drawings, but it should be understood that these embodiments are not intended to limit the present invention, and those skilled in the art should understand that functional, methodological, or structural equivalents or substitutions made by these embodiments are within the scope of the present invention.
Before describing in detail the various embodiments of the present invention, the core inventive concepts of the present invention are summarized and described in detail by the following several embodiments.
The invention provides a distortion null space planning design generation method and system based on a generation countermeasure network, which adopt a deep convolution neural network framework, can effectively utilize artificial intelligence technology and generate a proper design scheme according to an expected form of a user's distortion null space planning.
The first embodiment is as follows:
referring to fig. 1 to 3, the present embodiment discloses a method for generating a null space plan design based on generation of an anti-confrontation network (hereinafter referred to as "method").
Specifically, as shown in fig. 1, the method disclosed in this embodiment mainly includes the following steps:
step S1: establishing a data set of a real distortion zero space planning design scheme, and labeling an expected label for each design scheme in the data set.
Specifically, a plurality of design solutions are generated for the distorted null space, a data set of the design solutions is formed, and each design solution is labeled with a desired label, wherein the city remaining space/distorted null space (leftover space) is a conceptual description of a scattered and unorganized space in a block, including an intermediate space between buildings, a corner space between roads and bridges, an obsolete space with unknown purposes, an undesigned redundant space, and the like.
Then, step S2 is executed: and constructing a generation network and a discrimination network based on the deep convolution generation countermeasure network.
In particular, a Generative Adaptive Network (GAN) has been a popular research direction in the artificial intelligence society, and the basic idea of GAN is derived from two-person zero-sum game in game theory, which is composed of a generator and a discriminator, and is trained by means of antagonistic learning to estimate potential distribution of data samples and generate new data samples. The GAN is widely researched in the fields of image and visual calculation, voice and language processing, information security, chess games and the like, and has a huge application prospect. The Deep Convolutional antagonistic generation network DCGAN (Deep Convolutional adaptive Networks) is a combination of Convolutional neural Networks and antagonistic Networks. Deep convolution generated countermeasure network (DCGAN) belongs to a generation model network under unsupervised learning, and more people begin to pay attention to it due to its strong generation expression capability. The deep convolution generation countermeasure Network is formed by adding a convolution neural Network which is proved to have strong image expression capability on the basis of a generation countermeasure Network (GAN) and combining the generation countermeasure Network and the GAN, the architectures of the generation countermeasure Network and the GAN are basically the same, only the deep convolution generation countermeasure Network replaces a generator and a discriminator in a common generation countermeasure Network with two convolution neural networks which are optimized and improved, so that images can be generated better and true and false images can be classified, after the improved convolution neural Network is added, the whole generation countermeasure Network becomes easier to train, potential distribution of data samples can be estimated better, and new data samples can be generated.
The meaning and loss of a generator and a discriminator in the deep convolution generation countermeasure network are completely consistent with those of the originally generated countermeasure network, the structure of the discriminator is a convolution neural network, the input of the discriminator is an image, the input image is convolved by a plurality of layers to obtain a convolution characteristic, and the output is the probability that the image is a real image. For the generator, the whole structure is a convolution-like neural network structure based on the convolution neural network. Unlike a common neural network, the input of the generator is one-dimensional random noise instead of an image, and the conv convolution layer operation is also convolution, namely deconvolution, of a micro step instead of a common convolution operation.
And generating a generation network constructed by the confrontation network based on the deep convolution as a deep deconvolution neural network, and constructing a discrimination network as a deep convolution neural network. Convolutional Neural Networks (CNN) are a class of feed forward Neural Networks (fed forward Neural Networks) that contain convolution computations and have a deep structure, and are one of the representative algorithms for deep learning (deep learning). Convolutional Neural Networks have a feature learning (rendering) capability, and can perform Shift-Invariant classification (Shift-Invariant classification) on input information according to a hierarchical structure thereof, and are therefore also called "Shift-Invariant Artificial Neural Networks (SIANN)".
The deep convolutional neural network mainly comprises an input layer, a convolutional layer, an activation function, a pooling layer, a full-link layer and an output layer. The function of the convolution layer is to extract the characteristics of input data, the convolution layer internally comprises a plurality of convolution kernels, and each element forming the convolution kernels corresponds to a weight coefficient and a deviation quantity (bias vector), and is similar to a neuron (neuron) of a feedforward neural network. Each neuron in the convolution layer is connected to a plurality of neurons in a closely located region in the previous layer, the size of which depends on the size of the convolution kernel, and is referred to in the literature as the "receptive field", which means a field analogous to that of visual cortical cells. When the convolution kernel works, the convolution kernel regularly sweeps the input characteristics, matrix element multiplication summation is carried out on the input characteristics in the receptive field, and deviation amount is superposed. After the feature extraction is performed on the convolutional layer, the output feature map is transmitted to the pooling layer for feature selection and information filtering. The pooling layer contains a pre-set pooling function whose function is to replace the result of a single point in the feature map with the feature map statistics of its neighboring regions. The step of selecting the pooling area by the pooling layer is the same as the step of scanning the characteristic diagram by the convolution kernel, and the pooling size, the step length and the filling are controlled. The fully-connected layer in the convolutional neural network is equivalent to the hidden layer in the traditional feedforward neural network. The fully-connected layer is located at the last part of the hidden layer of the convolutional neural network and only signals are transmitted to other fully-connected layers. The feature map loses spatial topology in the fully connected layer, is expanded into vectors and passes through the excitation function. The convolutional neural network is usually a fully-connected layer upstream of the output layer, and thus has the same structure and operation principle as the output layer in the conventional feedforward neural network. For the image classification problem, the output layer outputs the classification label using a logistic function or a normalized exponential function (softmax function). In the object detection problem, the output layer may be designed to output the center coordinates, size, and classification of the object. In the image semantic segmentation, the output layer directly outputs the classification result of each pixel. The deconvolution is the reverse operation of the convolution operation, for a normal convolution operation, an image is obtained before convolution, and the features of the image can be obtained after convolution.
Then, referring to fig. 2, step S3 is performed: and generating a true and false judgment result of the first design scheme based on the generation network and the discrimination network according to the data set and the expected label.
Wherein, step S3 specifically includes the following contents:
step S31: inputting the expected label into the generation network, and outputting a corresponding first design scheme;
step S32: and simultaneously inputting the real design scheme, the first design scheme and the expected label corresponding to the first design scheme in the data set into the discrimination network, and outputting a true and false judgment result of the first design scheme.
Then, step S4 is executed: and generating a true and false judgment result of the random design scheme based on the discrimination network according to the data set and the expected label.
Wherein, step S4 specifically includes the following contents:
and simultaneously inputting a certain expected label in the data set, a real design scheme corresponding to the expected label and a random real design scheme into the discrimination network, and outputting a true and false judgment result of the random design scheme.
Then, referring to fig. 3, step S5 is performed: alternately and iteratively training the discrimination network and the generation network;
the main training process of the deep convolution generation countermeasure network is the same as that of the common generation countermeasure network, namely the countermeasure game process of the generator and the discriminator, the generator needs to be continuously trained and updated to enable the generated image to be more difficult to be distinguished by the discriminator, and the discriminator needs to be continuously trained to enable the false artificial image generated by the generator to be better separated from the real natural image.
Wherein, step S5 specifically includes the following contents:
and step S51, setting a loss function of the discriminant network and the generated network during training.
Specifically, the loss function of the discrimination network during training is set as:
wherein D represents a discrimination network, G represents a generation network, and PdataRepresents the distribution, P, of the data set x in step S1zRepresenting the distribution of the random noise signal z,to the expectation of the real design corresponding to the input expectation label y,to output the expectation of a random real design when the input expectation label is y,the desire to generate a design for the network for output when the input desired label is y.
The loss function of the generated network during training is set as follows:
step S52: and updating the parameters of the discrimination network and the generation network.
Specifically, the parameters of the discrimination network are updated:
updating parameters of the generation network:
finally, step S6 is executed: and receiving a desired label form submitted by a user, and generating a second design scheme corresponding to the desired label form by using the trained generation network.
Example two:
in combination with the method for generating a null-space plan design based on generation of an antagonistic network disclosed in the first embodiment, the present embodiment discloses a specific implementation example of a null-space plan design generation system (hereinafter referred to as "system") based on generation of an antagonistic network.
Referring to fig. 4, the system includes:
the data set generation module 11: establishing a data set of a real distortion null space planning design scheme, and labeling an expected label for each design scheme in the data set;
the network construction module 12: generating a confrontation network structure generation network and a discrimination network based on the deep convolution;
the first result generation module 13: generating a true and false judgment result of a first design scheme based on the generation network and the discrimination network according to the data set and the expected label;
the random result generation module 14: generating a true and false judgment result of a random design scheme based on the discrimination network according to the data set and the expected label;
the network training module 15: alternately and iteratively training the discrimination network and the generation network;
the second scenario generation module 16: and receiving a desired label form submitted by a user, and generating a second design scheme corresponding to the desired label form by using the trained generation network.
Specifically, in the data set generating module 11, a plurality of design solutions are generated for the distorted null space, and a data set of the design solutions is composed, and each design solution is labeled with a desired label, wherein the city remaining space/distorted null space (leftover space) is a conceptual description of a scattered and unorganized space in a block, including an intermediate space between buildings, a corner space between roads and bridges, a disused space with unknown use, an un-designed redundant space, and the like.
Specifically, in the network construction module 12, the generation network constructed by the countermeasure network based on the deep convolution is a deep deconvolution neural network, and the constructed discrimination network is a deep convolution neural network. The deconvolution is the reverse operation of the convolution operation, for a normal convolution operation, an image is obtained before convolution, and the features of the image can be obtained after convolution.
Specifically, the first result generation module 13 includes:
first pattern generation unit 131: inputting the expected label into the generation network, and outputting a corresponding first design scheme;
the first determination result generation unit 132: and simultaneously inputting the real design scheme, the first design scheme and the expected label corresponding to the first design scheme in the data set into the discrimination network, and outputting a true and false judgment result of the first design scheme.
Specifically, in the random result generation module 14, a desired label in the data set, a real design plan corresponding to the desired label, and a random real design plan are simultaneously input to the discrimination network, and a result of determining whether the random design plan is true or false is output.
Specifically, the network training module 15 includes:
function setting section 151: setting a loss function of the discrimination network and the generation network during training;
parameter updating unit 152: and updating the parameters of the discrimination network and the generation network.
Please refer to the first embodiment, which will not be described herein again, for a system for generating a null space planning design based on generation of an antagonistic network and a technical solution of the same parts in a method for generating a null space planning design based on generation of an antagonistic network disclosed in the first embodiment.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
In summary, the method has the advantages that the artificial intelligence technology is effectively utilized, the deep convolutional neural network framework is adopted, the teratocardiophore is planned and designed based on the mode of generating the countermeasure network, and a proper design scheme is generated according to an expected form of the user to the teratocardiophore planning. Because spatial structure is irregular, introduce the distortion space planning design with artificial intelligence technique, to practicing thrift the human cost and have the significance, can save designer's human cost on the one hand, on the other hand can save the demand side again and seek designer's time cost, convenient, swift demand side's demand simultaneously.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A distortion null space planning design generation method based on generation of an antagonistic network is characterized by comprising the following steps:
a data set generation step: establishing a data set of a real distortion null space planning design scheme, and labeling an expected label for each design scheme in the data set;
a network construction step: generating a confrontation network structure generation network and a discrimination network based on the deep convolution;
a first result generation step: generating a true and false judgment result of a first design scheme based on the generation network and the discrimination network according to the data set and the expected label;
a random result generation step: generating a true and false judgment result of a random design scheme based on the discrimination network according to the data set and the expected label;
network training: alternately and iteratively training the discrimination network and the generation network;
a second scheme generation step: and receiving a desired label form submitted by a user, and generating a second design scheme corresponding to the desired label form by using the trained generation network.
2. The method according to claim 1, wherein the generation network is a deep deconvolution neural network, and the discrimination network is a deep convolution neural network.
3. The method according to claim 1, wherein the first result generating step comprises:
a first scheme generation step: inputting the expected label into the generation network, and outputting a corresponding first design scheme;
a first judgment result generation step: and simultaneously inputting the real design scheme, the first design scheme and the expected label corresponding to the first design scheme in the data set into the discrimination network, and outputting a true and false judgment result of the first design scheme.
4. The method according to claim 1, wherein the random result generating step comprises:
and simultaneously inputting a certain expected label in the data set, a real design scheme corresponding to the expected label and a random real design scheme into the discrimination network, and outputting a true and false judgment result of the random design scheme.
5. The method according to claim 1, wherein the network training step comprises:
function setting step: setting a loss function of the discrimination network and the generation network during training;
and (3) updating parameters: and updating the parameters of the discrimination network and the generation network.
6. A distortion null space planning design generation system based on generation of an antagonistic network is characterized by comprising:
a dataset generation module: establishing a data set of a real distortion null space planning design scheme, and labeling an expected label for each design scheme in the data set;
a network construction module: generating a confrontation network structure generation network and a discrimination network based on the deep convolution;
a first result generation module: generating a true and false judgment result of a first design scheme based on the generation network and the discrimination network according to the data set and the expected label;
a random result generation module: generating a true and false judgment result of a random design scheme based on the discrimination network according to the data set and the expected label;
a network training module: alternately and iteratively training the discrimination network and the generation network;
a second scenario generation module: and receiving a desired label form submitted by a user, and generating a second design scheme corresponding to the desired label form by using the trained generation network.
7. The teratocardiographic design generation system of claim 6, wherein the generation network is a deep deconvolution neural network and the discrimination network is a deep convolution neural network.
8. The teratocardiology planning design generating system of claim 6, wherein the first result generating module comprises:
a first scheme generation unit: inputting the expected label into the generation network, and outputting a corresponding first design scheme;
a first determination result generation unit: and simultaneously inputting the real design scheme, the first design scheme and the expected label corresponding to the first design scheme in the data set into the discrimination network, and outputting a true and false judgment result of the first design scheme.
9. The teratocardiology space planning design generating system of claim 6, wherein the random result generating module comprises:
and simultaneously inputting a certain expected label in the data set, a real design scheme corresponding to the expected label and a random real design scheme into the discrimination network, and outputting a true and false judgment result of the random design scheme.
10. The teratocardiology planning design generating system of claim 6, wherein the network training module comprises:
a function setting unit: setting a loss function of the discrimination network and the generation network during training;
a parameter updating unit: and updating the parameters of the discrimination network and the generation network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011321083.6A CN112329116A (en) | 2020-11-23 | 2020-11-23 | Distortion zero space planning design generation method and system based on generation of countermeasure network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011321083.6A CN112329116A (en) | 2020-11-23 | 2020-11-23 | Distortion zero space planning design generation method and system based on generation of countermeasure network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112329116A true CN112329116A (en) | 2021-02-05 |
Family
ID=74321409
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011321083.6A Pending CN112329116A (en) | 2020-11-23 | 2020-11-23 | Distortion zero space planning design generation method and system based on generation of countermeasure network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112329116A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20000059246A (en) * | 2000-07-26 | 2000-10-05 | 조근식 | System and method for spatial layout planning over the web |
CN107220600A (en) * | 2017-05-17 | 2017-09-29 | 清华大学深圳研究生院 | A kind of Picture Generation Method and generation confrontation network based on deep learning |
CN108829855A (en) * | 2018-06-21 | 2018-11-16 | 山东大学 | It is worn based on the clothing that condition generates confrontation network and takes recommended method, system and medium |
CN109886200A (en) * | 2019-02-22 | 2019-06-14 | 南京邮电大学 | A kind of unmanned lane line detection method based on production confrontation network |
CN111553587A (en) * | 2020-04-26 | 2020-08-18 | 中国电力科学研究院有限公司 | New energy scene generation method and system based on confrontation learning model |
WO2020172838A1 (en) * | 2019-02-26 | 2020-09-03 | 长沙理工大学 | Image classification method for improvement of auxiliary classifier gan |
-
2020
- 2020-11-23 CN CN202011321083.6A patent/CN112329116A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20000059246A (en) * | 2000-07-26 | 2000-10-05 | 조근식 | System and method for spatial layout planning over the web |
CN107220600A (en) * | 2017-05-17 | 2017-09-29 | 清华大学深圳研究生院 | A kind of Picture Generation Method and generation confrontation network based on deep learning |
CN108829855A (en) * | 2018-06-21 | 2018-11-16 | 山东大学 | It is worn based on the clothing that condition generates confrontation network and takes recommended method, system and medium |
CN109886200A (en) * | 2019-02-22 | 2019-06-14 | 南京邮电大学 | A kind of unmanned lane line detection method based on production confrontation network |
WO2020172838A1 (en) * | 2019-02-26 | 2020-09-03 | 长沙理工大学 | Image classification method for improvement of auxiliary classifier gan |
CN111553587A (en) * | 2020-04-26 | 2020-08-18 | 中国电力科学研究院有限公司 | New energy scene generation method and system based on confrontation learning model |
Non-Patent Citations (2)
Title |
---|
何平;张万发;罗萌;: "基于生成对抗网络的遥感图像居民区分割", 传感器与微***, vol. 39, no. 02, pages 113 - 116 * |
刘跃中等: "基于条件生成对抗网络的城市设计研究", 建筑学报, no. 9, pages 108 - 113 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109887282B (en) | Road network traffic flow prediction method based on hierarchical timing diagram convolutional network | |
CN111310773B (en) | Efficient license plate positioning method of convolutional neural network | |
Shi et al. | Road detection from remote sensing images by generative adversarial networks | |
CN109508360B (en) | Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton | |
CN104809187B (en) | A kind of indoor scene semanteme marking method based on RGB D data | |
CN110503112A (en) | A kind of small target deteection of Enhanced feature study and recognition methods | |
CN108537824B (en) | Feature map enhanced network structure optimization method based on alternating deconvolution and convolution | |
Zhu et al. | ED-Net: Automatic building extraction from high-resolution aerial images with boundary information | |
CN108805345A (en) | A kind of crime space-time Risk Forecast Method based on depth convolutional neural networks model | |
CN113159067A (en) | Fine-grained image identification method and device based on multi-grained local feature soft association aggregation | |
Shu et al. | CFSA-Net: Efficient Large-Scale Point Cloud Semantic Segmentation Based on Cross-Fusion Self-Attention. | |
CN113052254A (en) | Multi-attention ghost residual fusion classification model and classification method thereof | |
Xu et al. | BANet: A balanced atrous net improved from SSD for autonomous driving in smart transportation | |
CN110084136A (en) | Context based on super-pixel CRF model optimizes indoor scene semanteme marking method | |
Hiramatsu et al. | Cell image segmentation by integrating multiple CNNs | |
Sanchez-Valdes et al. | Linguistic description about circular structures of the Mars’ surface | |
Luo | Online design of green urban garden landscape based on machine learning and computer simulation technology | |
Singh et al. | Iml-gcn: Improved multi-label graph convolutional network for efficient yet precise image classification | |
Chen et al. | Research on object detection algorithm based on multilayer information fusion | |
CN112329116A (en) | Distortion zero space planning design generation method and system based on generation of countermeasure network | |
CN103247076B (en) | A kind of three-dimensional model synthetic method based on global optimization | |
CN111190952B (en) | Method for extracting and persistence of multi-scale features of city portrait based on image pyramid | |
Dong et al. | SiameseDenseU‐Net‐based Semantic Segmentation of Urban Remote Sensing Images | |
Song | The Fine Design Strategy of Urban Streets Using Deep Learning with the Assistance of the Internet of Things | |
Fei et al. | A GNN Architecture With Local and Global-Attention Feature for Image Classification |
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 |