Automatic Generation Method, Device, Computer Equipment and Storage Medium for Housing Type Layout

Abstract

The present application discloses a method, apparatus, computer device and storage medium for automatically generating a housing layout, which relates to the field of computer application technologies. The method includes obtaining the image features of a housing type, performing autoregressive processing on the image features to obtain multiple layout results corresponding to the housing type. The layout results include the structural layout relationships representing the housing type. Selecting a target layout result from the multiple layout results. On the one hand, since autoregressive processing can generate countless different prediction results for a fixed input, this technical solution utilizes this feature and combines home design with deep learning, making the generated layout results more diverse and robust, and providing more housing layout references for users. On the other hand, further screening the output layout results makes the finally output target layout result closer to the user's demand trend for the housing layout, improving the success rate and efficiency of using the target layout result.

Description

Technical Field

[0001] This application relates to the field of computer application technologies, and in particular, to a method, apparatus, computer device, and storage medium for automatically generating house layouts. Background Art

[0002] In recent years, with the continuous integration of home design and computer technology, the intelligent design of indoor homes has received extensive attention. Generally, furniture design layouts are based on fixed design templates, resulting in very limited and single furniture layout results provided for users. Summary of the Invention

[0003] The purpose of the embodiments of this application is to propose a method for automatically generating house layouts to solve the problem of single furniture layout results.

[0004] To solve the above technical problems, the embodiments of this application provide a method for automatically generating house layouts, including the following steps:

[0005] Obtain the image features of the house type;

[0006] Perform autoregressive processing on the image features to obtain multiple layout results corresponding to the house type, where the layout results include the structural layout relationships representing the house type;

[0007] Select a target layout result from the multiple layout results.

[0008] In some embodiments, performing autoregressive processing on the image features to obtain multiple layout results corresponding to the house type includes:

[0009] Obtain multiple preset initial furniture sequences, where the number of initial furniture sequences is the same as the number of layout results;

[0010] Input the image features and each initial furniture sequence into a preset autoregressive model for prediction to output each predicted target furniture sequence, where each value in the target furniture sequence represents furniture attribute information;

[0011] Output the layout results corresponding to each house type according to each target furniture sequence and the image features.

[0012] In some embodiments, obtaining the image features of the house type includes:

[0013] Obtain the house type image;

[0014] Extract the image features of the house type image through a preset residual network.

[0015] In some embodiments, after outputting each predicted target furniture sequence, the method further includes:

[0016] Obtain the customized furniture attribute information and update the customized furniture attribute information into the target sequence.

[0017] In some embodiments, screening out the target layout result from multiple layout results includes:

[0018] Obtain the scoring value of each layout result;

[0019] Filter the scoring values according to a preset scoring threshold;

[0020] Take the layout result corresponding to the filtered scoring value as the target layout result.

[0021] In some embodiments, obtaining the scoring value of each layout result includes:

[0022] Obtain multiple initial scoring values of each layout result;

[0023] Sum up the multiple initial scoring values of each layout result to obtain the scoring value of each layout result.

[0024] To solve the above technical problems, an embodiment of the present application further provides an automatic generation device for housing layout, including:

[0025] An acquisition module, configured to acquire the image features of the housing type;

[0026] An autoregressive processing module, configured to perform autoregressive processing on the image features to obtain multiple layout results corresponding to the housing type, where the layout results include the structural layout relationship representing the housing type;

[0027] A screening module, configured to screen out the target layout result from multiple layout results.

[0028] In some embodiments, the autoregressive processing module includes:

[0029] An acquisition unit, configured to acquire multiple preset initial furniture sequences, where the number of initial furniture sequences is the same as the number of layout results;

[0030] A regression unit, configured to input the image features and each initial furniture sequence into a preset autoregressive model for prediction to output each predicted target furniture sequence, where each value in the target furniture sequence represents furniture attribute information;

[0031] A layout unit, configured to output the layout result corresponding to each housing type according to each target furniture sequence and the image features.

[0032] In some embodiments, the acquisition module includes:

[0033] An image acquisition unit, configured to acquire the housing type image;

[0034] An extraction unit for extracting image features of a housing type image through a preset residual network.

[0035] In some embodiments, the automatic generation device for housing type layout further includes:

[0036] A custom module for obtaining custom furniture attribute information and updating the custom furniture attribute information to the target sequence.

[0037] In some embodiments, the screening module includes:

[0038] A scoring value acquisition unit for acquiring the scoring value of each layout result;

[0039] A filtering unit for filtering the scoring value according to a preset scoring threshold;

[0040] A target layout unit for using the layout result corresponding to the filtered scoring value as the target layout result.

[0041] In some embodiments, the scoring value acquisition unit includes:

[0042] An acquisition subunit for acquiring multiple initial scoring values of each layout result;

[0043] A calculation subunit for summing up the multiple initial scoring values of each layout result to obtain the scoring value of each layout result.

[0044] To solve the above technical problems, an embodiment of the present application further provides a computer device, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps of the above automatic generation method for housing type layout are implemented.

[0045] To solve the above technical problems, an embodiment of the present application further provides a computer-readable storage medium. A computer program is stored on the computer-readable storage medium, and when the computer program is executed by the processor, the steps of the above automatic generation method for housing type layout are implemented.

[0046] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects:

[0047] By obtaining the image features of the housing type, performing autoregressive processing on the image features to obtain various layout results corresponding to the housing type, where the layout results include the structural layout relationship representing the housing type, and screening out the target layout result from the various layout results. On the one hand, since autoregressive processing is a mathematical model based on probability prediction of time series and has the characteristic of being able to output countless prediction sequences from the prediction process, that is, this characteristic can generate countless different prediction results for a fixed input. Therefore, this technical solution uses this characteristic and combines the way of home design and deep learning, making the generated layout results more diverse and robust, providing more housing type layout references for users. On the other hand, further screening the output layout results makes the finally output target layout result closer to the user's demand trend for the housing type layout, improving the success rate and efficiency of using the target layout result. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] To more clearly illustrate the solutions in this application, the following will briefly introduce the drawings required for the description of the embodiments of this application. Obviously, the drawings in the following description are some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0049] Figure 1 is an exemplary system architecture diagram to which this application can be applied;

[0050] Figure 2 is a flowchart of an embodiment of the method for automatically generating the housing type layout provided by this application;

[0051] Figure 3 is a schematic structural diagram of the deep residual network provided by this application;

[0052] Figure 4 is a schematic flowchart of the housing type image after autoregressive processing provided by the embodiment of this application;

[0053] Figure 5 is a schematic diagram of another embodiment of the method for automatically generating the housing type layout provided by this application;

[0054] Figure 6 is a schematic diagram of the scenario for automatically generating the housing type layout provided by this application;

[0055] Figure 7 is a schematic structural diagram of an embodiment of the device for automatically generating the housing type layout provided by this application;

[0056] Figure 8 is a schematic structural diagram of an embodiment of the computer device provided by this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs; the terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification, claims and drawings of this application are intended to cover non-exclusive inclusion. The terms "first", "second", etc. in the specification, claims or drawings of this application are used to distinguish different objects and not to describe a specific order.

[0058] Reference to "embodiment" herein means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of this application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0059] To enable those skilled in the art to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0060] As Figure 1 shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0061] Users can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0062] The terminal devices 101, 102, and 103 can be various electronic devices with a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, and so on.

[0063] The server 105 can be a server that provides various services, such as a background server that supports the pages displayed on the terminal devices 101, 102, and 103.

[0064] It should be noted that the method for automatically generating a housing layout provided in the embodiments of the present application is generally executed by a server / terminal. Terminal device Correspondingly, the device for automatically generating a housing layout is generally set in a server / terminal device.

[0065] It should be understood that Figure 1 the numbers of terminal devices, networks, and servers in

[0066] are merely illustrative. According to the implementation requirements, there can be any number of terminal devices, networks, and servers. Figure 2 Continuing to refer to

[0067] FIG. 18 shows a flowchart of an embodiment of the method for automatically generating a housing layout of the present application. The method for automatically generating a housing layout includes the following steps:

[0068] S201: Obtain the image features of the housing type.

[0069] In some embodiments, obtaining the image features of the housing type includes:

[0070] Obtain a housing type image;

[0071] Extract the image features of the housing type image through a preset residual network.

[0072] Specifically, it can be that the user inputs a house type image in the terminal device; it can be obtained by querying from a data database that pre-stores the house type image according to the user information; it can be that the user obtains the house type image uploaded by the user through wireless network / Bluetooth / 4G / 5G, etc. The acquisition method is not limited in this application. The house type image can be a two-dimensional image or a three-dimensional image, can be an RGB image, or can be a grayscale image. The image type of the house type image is not limited here.

[0073] For the extraction of the image features of the house type image, the Histogram of Oriented Gradient (HOG), Local Binary Pattern (LBP), or deep learning neural network algorithm can be used, which is not limited here.

[0074] Furthermore, the embodiment of this application uses a deep learning neural network algorithm to extract image features. For example, in the embodiment of this application, a deep residual network is used to extract features from the house type image. The deep residual network can not only extract deeper semantic information by increasing the depth of the neural network to improve the accuracy of feature extraction, but also uses skip connections to alleviate the problem of gradient disappearance caused by increasing the depth in the deep neural network.

[0075] Please refer to Figure 3 , Figure 3 which is a schematic structural diagram of the deep residual network provided by this application. As Figure 3 shown, the deep residual network includes 5 residual modules, and each residual module consists of two 3×3 convolutions. Specifically, a 7×96 house type image is input into the deep residual network. After convolution with 3×3×64, the size of the feature map obtained by convolution is halved to obtain a 4×48×64 feature map. After the 4×48×64 feature map is subjected to maximum pooling processing with 3×3×64, the size of the feature map obtained by pooling is halved to obtain a 2×24×64 feature map. Then, the 2×24×64 feature map is processed through five residual modules to obtain a 2×24×64 feature map, and then further through global average pooling processing to finally obtain the image features of the house type. Among them, the convolution depth in the embodiment of this application is 64, and the size of the feature map remains unchanged during the residual processing.

[0076] It should be noted that the specific dimensions of the input image and the number of residual modules are hyperparameters and are adjusted according to the actual situation.

[0077] In the embodiment of this application, the image feature is the feature vector of the house type. The process of obtaining the feature vector x vec by processing the house type image x through the deep residual network can be expressed as xvec = ResNet(x). The feature vector may include the housing type category, the azimuth size of the housing type category, and also furniture attribute information such as furniture category, the quantity of furniture categories, the size of furniture categories, the furniture placement position and orientation, etc. For example, the housing type category can be a living room, a bedroom, a kitchen, a bathroom, etc., and the furniture category can be a TV cabinet, a sofa, a toilet, a bed, a washbasin, etc., which are not limited here.

[0078] S202: Perform autoregressive processing on the image features to obtain multiple layout results corresponding to the housing type, where the layout results include the structural layout relationship representing the housing type.

[0079] Among them, the autoregressive processing is implemented through an autoregressive model. Among them, the autoregressive model uses a Transformer (transformation model), and the Transformer is a model that uses an attention mechanism to improve the model training speed. The Transformer architecture has two input sequences, namely the input sequence of the encoder and the input sequence of the decoder. The structures of each encoder are the same but do not share the respective weights of the encoders. Each encoder includes a self-attention layer and a fully connected feed-forward network layer. Each decoder also includes a self-attention layer and a fully connected feed-forward network layer, and there is also an attention layer between these two layers.

[0080] In some embodiments, performing autoregressive processing on the image features to obtain multiple layout results corresponding to the housing type includes:

[0081] Obtain multiple preset initial furniture sequences, where the number of initial furniture sequences is the same as the number of layout results;

[0082] Input the image features and each initial furniture sequence into a preset autoregressive model for prediction to output each predicted target furniture sequence, where each value in the target furniture sequence represents furniture attribute information;

[0083] Output the layout results corresponding to each housing type according to each target furniture sequence and the image features.

[0084] In the embodiments of the present application, the image features serve as the input sequence of the encoder, and the initial furniture sequence serves as the input sequence of the decoder. Moreover, the initial furniture sequence is pre-cyclically predicted through an autoregressive model. Therefore, each initial furniture sequence can output a target furniture sequence after being processed by the autoregressive model. The number of set initial furniture sequences is the same as the number of output target furniture sequences. Each value in the target furniture sequence represents the predicted value of each furniture. For example, each value can be used to represent each furniture category, furniture size, etc. That is, the output target furniture sequence determines the structural layout relationship of the housing type. The structural layout relationship of the housing type reflects the size of each furniture category and the positional relationship in the housing type. Generate the layout result of the housing type according to the predicted target furniture sequence and the image features. The layout result includes the image of the layout of each furniture category in the housing type, and the rendering is as shown in Figure 6 shown.

[0085] Specifically, input the image features x vec and the initial furniture sequence v seq into the autoregressive model (Transformer) to obtain the probability distribution matrix of the next value in the initial furniture sequence, denoted as Transformer(x vec , v seq ). Perform a sampling operation (sample) on the probability distribution matrix of the next value, and denote the sampled value as v n . The entire process of obtaining v n can be expressed as v n =sample(Transformer(x vec , v seq ). Among them, the sampling operation can be to take the value with the largest probability in the probability distribution matrix, or obtain the optimal value through Nucleus Sampling. Further, add v n to the end of the initial furniture sequence to update the initial furniture sequence, and continue to use the updated initial furniture sequence according to v n =sample(Transformer(x vec , v seq ) to obtain the next v n value until the prediction stops when all values of the entire initial furniture sequence have been traversed and predicted. At this time, the target furniture sequence is output. Among them, the preset value of each value when each initial furniture sequence is input into the autoregressive model can be all 0, 1,..., n - 1, n, where n is a positive integer. For example, the first initial furniture sequence = [0, 0, 0, 0, 0], the second initial furniture sequence = [1, 1, 1, 1, 1],..., the nth initial furniture sequence = [n, n, n, n, n].

[0086] For example, when the autoregressive model needs to predict the first value of the first initial furniture sequence, the probability distribution of this value is obtained through autoregressive processing. A preset sampling operation is adopted. For example, the value with the largest probability in the probability distribution is taken as this value. Assume the first value is 1 and it is added to the end of the initial furniture sequence. At this time, the updated initial furniture sequence = [1, ]. The updated initial furniture sequence = [1, ] is continuously input into the autoregressive model, and then the sampling operation is performed to obtain the second value of the initial furniture sequence. Assume the second value is 3. At this time, the initial furniture sequence = [1, ] is updated to obtain the initial furniture sequence = [1, 3]... and so on in a loop. Finally, the updated values in the entire initial furniture sequence are obtained, that is, the target furniture sequence. For example, at this time, the target furniture sequence = [1, 3, 2, 4, 5, 1, 4, 5, 7, 8]. This initial furniture sequence represents a chair with a size of [3, 2] placed at the position of [4, 5], and a table with a size of [4, 5] placed at the position of [7, 8]. That is to say, the first 5 values represent the attribute information of a chair, and the last 5 values represent the attribute information of a table.

[0087] In the embodiments of the present application, each target furniture sequence can represent a layout result, that is, the layout scheme of the house type. The probability of the target furniture sequence can be calculated as the probability of a layout in the house type. The reliability of the layout result of the target furniture sequence can be determined by probability calculation for the finally output target furniture. The calculation formula is where p is the probability of the entire target furniture sequence; θ is the autoregressive model parameter; V seq represents the target furniture sequence here; v n represents the nth value in the target furniture sequence. It can be seen from the above formula that the probabilities of each value in the target sequence are multiplied to finally obtain the probability of the target furniture sequence. The greater the probability of the target furniture sequence, the more the output layout result conforms to the structural layout relationship of the current house type.

[0088] In some embodiments, after each predicted target furniture sequence is output, the method further includes:

[0089] Obtain the custom furniture attribute information and use the custom furniture attribute information as the new furniture attribute information.

[0090] Among them, the furniture attribute information includes furniture category, furniture size, and the placement position of the furniture in the housing type. For example, in the bathroom scene, the furniture category may include washbasins, bathtubs, toilets, etc., and may be the size of the washbasin and the position of the washbasin in the housing type. The user can manipulate the furniture attribute information of the prediction sequence during the housing layout design process, so that the layout result of the obtained housing type is closer to the user's needs, and at the same time, the interactivity of the housing layout design is improved. For example, in the furniture layout process of the bedroom, the user can arbitrarily select the size and orientation of the bed as the new furniture attribute information, and after autoregressive prediction, output the placement position of the bed in the current housing type.

[0091] Continue to refer to Figure 4 , Figure 4 is a schematic flow chart of the autoregressive processing of the housing type image provided by the embodiment of the present application. As Figure 4 shown, obtain the housing type image input by the user, extract the features of the housing type image through the feature extraction network to obtain image features, and generate a prediction sequence including feature vectors through the autoregressive model, so that the furniture attribute information in the feature vectors of the prediction sequence reflects the layout results of each furniture category of the housing type. For example, the furniture attribute information included in the prediction sequence is a two-dimensional vector in the actual scene, and the prediction sequence is a one-dimensional vector in the embodiment of the present application. Therefore, it is necessary to split the two-dimensional vector. For example, split the length and width of the furniture size into length and width. The one-dimensional vector can represent the furniture category, the furniture size, and the orientation of the furniture placement. At the same time, the user can adjust the prediction sequence by customizing and controlling the furniture attribute information of the feature vectors in the terminal device and generate a new layout result.

[0092] Furthermore, record the behavior information such as the user's selection and adjustment of the furniture attribute information, and then transform it into a series of housing design solutions available for the autoregressive model to learn, and finally realize the iterative optimization of the autoregressive model.

[0093] S203: Screen out the target layout result from multiple layout results.

[0094] Since the number of layout results obtained through autoregressive prediction is large, and not every layout result is reasonable and effective in terms of the layout relationship with the housing type structure. Therefore, it is necessary to screen each layout result to output the layout results with relatively high rankings as the target layout results, so as to improve the effectiveness of the housing layout plan.

[0095] It should be noted that in the embodiment of the present application, the screening can be performed first according to the probability of the target furniture sequence. For example, the target furniture sequence greater than the preset probability threshold is used as the layout result, and then through the following screening method, the target layout result is obtained.

[0096] In some embodiments, screening out the target layout result from multiple layout results includes:

[0097] Obtaining the scoring value of each layout result;

[0098] Filtering the scoring values according to a preset scoring threshold;

[0099] Taking the layout result corresponding to the filtered scoring value as the target layout result.

[0100] In the embodiments of the present application, obtaining the scoring value of each layout result includes:

[0101] Obtaining multiple initial scoring values of each layout result;

[0102] Summing up the multiple initial scoring values of each layout result to obtain the scoring value of each layout result.

[0103] Specifically, each rule can be a function and input a corresponding score. The scoring value output by each layout result in each rule is the initial scoring value. Suppose a certain layout result is layout i , and its score under a certain rule rule j is score i，j = rule j (layout i ). The total score of its layout layout i is: score i = ∑ j rule j (layout i ), that is, summing up multiple initial scoring values and taking the summation result as the scoring value.

[0104] Furthermore, each layout result can be evaluated through a pre-scoring system to obtain the scoring value of each layout result. Among them, the scoring system consists of a series of scoring rules, and each rule is preset with a corresponding score value. Therefore, it is queried whether each layout result satisfies each scoring rule in the scoring system, and by counting the score values corresponding to whether the rules are satisfied or not, the final scoring value of each layout result is obtained.

[0105] It should be noted that the user can adjust the scoring rules in the scoring system. For example, if a certain rule is unreasonable, it can be selected to be deleted.

[0106] Further, the preset scoring threshold can be set according to the actual scenario. Compare the scoring values of each layout result with the scoring threshold, filter out the layout results with scores less than the scoring threshold, and obtain the layout results greater than the scoring threshold. For example, if the scoring threshold is 90, the layout results with scoring values greater than or equal to 90 of the layout results are used as the target results.

[0107] Continue to refer to Figure 5 , Figure 5 which is a schematic diagram of another embodiment of the method for automatically generating the household layout provided by the present application. Input multiple furniture layout results into the scoring system in parallel. Score the layout plans in the furniture layout results through a series of different scoring rules, that is, obtain corresponding scores according to whether the scoring rules are satisfied, and count the total scores of each furniture layout result in all scoring rules. Sort according to the total scores of each layout result, for example, sort according to the scores from high to low, filter out the layout results with too low scores, and then output the final furniture layout result.

[0108] Screen the layout results through the scoring values, making the output target layout results more efficient and making the structural layout relationship of the household more reasonable.

[0109] Further, as Figure 6 shown, Figure 6 is a schematic diagram of the scenario for automatically generating the household layout provided by the present application. As Figure 6 shown, the first figure shown is the input household structure diagram, for example, the structure diagram of a bathroom, where gray represents the wall, white represents the room door, and black represents the floor. The user defines relevant parameters for the household structure diagram, for example, the parameter definition includes newly added furniture categories, furniture attribute information, etc. or adjusts relevant parameters. For example, Figure 6 the household structure diagram located below the first figure has newly added squares of different colors compared to the first figure. Taking the bathroom as an example, the user can add a washbasin, a toilet, a shower room, etc. to the household structure diagram. At the same time, the user can also adjust the values of the width and depth of the shower room, and further screen the layout results obtained through autoregressive processing to obtain the final layout results. As Figure 6 shown in the two household structure diagrams on the right, it not only includes the furniture categories input by the user, but also shows the placement position relationship of each furniture category in the current household.

[0110] By obtaining the image features of the house type, performing autoregressive processing on the image features to obtain various layout results corresponding to the house type, where the layout results include the structural layout relationship representing the house type, and screening out the target layout result from the various layout results. On the one hand, since the autoregressive processing is a mathematical model based on probability prediction time series and has the characteristic of being able to output countless prediction sequences during the prediction process, that is, this characteristic can generate countless different prediction results for a fixed input. Therefore, this technical solution utilizes this characteristic and combines the way of home design and deep learning, making the generated layout results more diverse and robust, providing more house type layout references for users. On the other hand, further screening the output layout results makes the finally output target layout result closer to the user's demand trend for the house type layout, improving the success rate and efficiency of using the target layout result.

[0111] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above various methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a Read-Only Memory (ROM), etc., or a Random Access Memory (RAM), etc.

[0112] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit and can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages do not necessarily have to be executed at the same moment, but can be executed at different moments, and their execution order does not necessarily have to be sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0113] Further reference Figure 7 to Figure 2 As an implementation of the method shown above, an embodiment of an automatic generation device for house type layout is provided in this application. This device embodiment corresponds to the method embodiment shown in Figure 2 and this device can be specifically applied to various electronic devices.

[0114] As Figure 7 shown, Figure 7A structural schematic diagram of an embodiment of the automatic generation device for housing layout provided by this application. Among them, the automatic generation device for housing layout further includes: an acquisition module 701, an autoregressive processing module 702, and a screening module 703. Among them:

[0115] The acquisition module 701 is used to acquire the image features of the housing type;

[0116] The autoregressive processing module 702 is used to perform autoregressive processing on the image features to obtain multiple layout results corresponding to the housing type. Among them, the layout results include the structural layout relationships representing the housing type;

[0117] The screening module 703 is used to screen out the target layout results from multiple layout results.

[0118] In some embodiments, the autoregressive processing module 702 includes:

[0119] An acquisition unit, used to acquire multiple preset initial furniture sequences, where the number of initial furniture sequences is the same as the number of layout results;

[0120] A regression unit, used to input the image features and each initial furniture sequence into a preset autoregressive model for prediction to output each predicted target furniture sequence, where each value in the target furniture sequence represents furniture attribute information;

[0121] A layout unit, used to output the layout results corresponding to each housing type according to each target furniture sequence and the image features.

[0122] In some embodiments, the acquisition module 701 includes:

[0123] An image acquisition unit, used to acquire the housing type image;

[0124] An extraction unit, used to extract the image features of the housing type image through a preset residual network.

[0125] In some embodiments, the automatic generation device for housing layout further includes:

[0126] A customization module, used to acquire customized furniture attribute information and update the customized furniture attribute information to the target sequence.

[0127] In some embodiments, the screening module 703 includes:

[0128] A scoring value acquisition unit, used to acquire the scoring value of each layout result;

[0129] A filtering unit, used to filter the scoring value according to a preset scoring threshold;

[0130] A target layout unit for using the layout result corresponding to the filtered score value as the target layout result.

[0131] In some embodiments, the score value acquisition unit includes:

[0132] An acquisition subunit for acquiring multiple initial score values of each layout result;

[0133] A calculation subunit for summing up the multiple initial score values of each layout result to obtain the score value of each layout result.

[0134] Regarding the automatic generation device for the housing layout in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0135] To solve the above technical problems, an embodiment of the present application also provides a computer device. For details, please refer to Figure 8 , Figure 8 which is the basic structural block diagram of the computer device in this embodiment.

[0136] The computer device 8 includes a memory 81, a processor 82, and a network interface 83 that are communicatively connected to each other through a system bus. It should be noted that only the computer device 8 with components 81 - 83 is shown in the figure, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be alternatively implemented. Among them, those skilled in the art of the present technology can understand that a computer device here is a device capable of automatically performing numerical calculations and / or information processing according to pre - set or stored instructions, and its hardware includes, but is not limited to, a microprocessor, an application - specific integrated circuit (ASIC), a field - programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0137] The computer device can be a desktop computer, a notebook, a palm computer, a cloud server, and other computing devices. The computer device can perform human - computer interaction with users through a keyboard, a mouse, a remote control, a touchpad, a voice - controlled device, etc.

[0138] The memory 81 includes at least one type of readable storage medium, which includes flash memory, hard disk, multimedia card, card-type memory (e.g., SD or D-form layout automatic generation memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 81 may be an internal storage unit of the computer device 8, such as the hard disk or memory of the computer device 8. In other embodiments, the memory 81 may also be an external storage device of the computer device 8, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc. equipped on the computer device 8. Of course, the memory 81 may also include both the internal storage unit and the external storage device of the computer device 8. In this embodiment, the memory 81 is generally used to store the operating system and various application software installed in the computer device 8, such as the program code of the method for automatic generation of the house type layout. In addition, the memory 81 may also be used to temporarily store various types of data that have been output or will be output.

[0139] In some embodiments, the processor 82 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 82 is generally used to control the overall operation of the computer device 8. In this embodiment, the processor 82 is used to run the program code stored in the memory 81 or process data, such as running the program code of the method for automatic generation of the house type layout.

[0140] The network interface 83 may include a wireless network interface or a wired network interface, and the network interface 83 is generally used to establish a communication connection between the computer device 8 and other electronic devices.

[0141] This application also provides another implementation manner, that is, to provide a computer-readable storage medium storing an automatic generation program for the house type layout, and the automatic generation program for the house type layout can be executed by at least one processor, so that the at least one processor executes the steps of the method for automatic generation of the house type layout as described above.

[0142] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases, the former is a better implementation method. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present application.

[0143] Obviously, the above-described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The accompanying drawings show the preferred embodiments of the present application, but do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosed content of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions described in the foregoing specific embodiments, or perform equivalent replacements for some of the technical features. Any equivalent structure directly or indirectly using the content of the specification and drawings of the present application in other related technical fields is equally within the scope of the patent protection of the present application.

Claims

1. An automatic generation method for apartment layout, characterized in that, the method includes: Obtaining the image features of the apartment layout; Performing autoregressive processing on the image features to obtain multiple layout results corresponding to the apartment layout, wherein the layout results include the structural layout relationship representing the apartment layout; Selecting a target layout result from the multiple layout results; The performing autoregressive processing on the image features to obtain multiple layout results corresponding to the apartment layout includes: Obtaining a plurality of preset initial furniture sequences, wherein the number of initial furniture sequences is the same as the number of layout results; Inputting the image features and each of the initial furniture sequences into a preset autoregressive model for prediction to output each predicted target furniture sequence, wherein each value in the target furniture sequence represents furniture attribute information; Outputting the layout results corresponding to each apartment layout according to each target furniture sequence and the image features; After outputting each predicted target furniture sequence, the method further includes: Obtaining custom furniture attribute information and updating the custom furniture attribute information to the target furniture sequence; The selecting a target layout result from the multiple layout results includes: Obtaining the scoring value of each layout result; Filtering the scoring value according to a preset scoring threshold; Taking the layout result corresponding to the filtered scoring value as the target layout result.

2. The automatic generation method for apartment layout according to claim 1, characterized in that, the obtaining the image features of the apartment layout includes: Obtaining an apartment layout image; Extracting the image features of the apartment layout image through a preset residual network.

3. The automatic generation method for apartment layout according to claim 1, characterized in that, the obtaining the scoring value of each layout result includes: Obtaining multiple initial scoring values of each layout result; Summing the multiple initial scoring values of each layout result to obtain the scoring value of each layout result.

4. An automatic generation device for apartment layout, characterized in that, the automatic generation device for apartment layout includes: An obtaining module, configured to obtain the image features of the apartment layout; An autoregressive processing module, configured to perform autoregressive processing on the image features to obtain multiple layout results corresponding to the apartment layout, wherein the layout results include the structural layout relationship representing the apartment layout; A screening module, configured to select a target layout result from the multiple layout results, including: Obtaining the scoring value of each layout result; Filtering the scoring value according to a preset scoring threshold; Taking the layout result corresponding to the filtered scoring value as the target layout result; The autoregressive processing module includes: An obtaining unit, configured to obtain a plurality of preset initial furniture sequences, wherein the number of initial furniture sequences is the same as the number of layout results; A regression unit for inputting the image features and each of the initial furniture sequences into a preset autoregressive model for prediction to output each predicted target furniture sequence, where each value in the target furniture sequence represents furniture attribute information; after outputting each predicted target furniture sequence, it further includes: obtaining custom furniture attribute information and updating the custom furniture attribute information to the target furniture sequence; A layout unit for outputting a layout result corresponding to each house type according to each target furniture sequence and the image features.

5. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the automatic generation method for house type layout according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium, characterized in that, a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the automatic generation method for house type layout according to any one of claims 1 to 3 are implemented.

Images