Intelligent bathroom decoration scheme generation method and device

Abstract

The invention provides a bathroom decoration scheme intelligent generation method and device. The method comprises the following steps: acquiring a decoration demand of a customer and a to-be-customized indoor space real-shot picture; a decoration demand is input into the trained bathroom design scheme to generate a model, and the bathroom design scheme is predicted through the model; and generating a rendering graph of the indoor space to be customized by using the bathroom design scheme and the real shot graph of the indoor space to be customized. According to the method, the design efficiency and quality are improved, and more personalized and targeted design experience is provided for customers; meanwhile, the cost is saved for enterprises, the market competitiveness is improved, and the time for demand analysis and budget evaluation in the traditional manual customization process is greatly shortened; the demand point of the customer can be more accurately grasped, and the situation that the design scheme does not conform to the expectation of the customer due to understanding deviation is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of interior decoration, and in particular to a method and device for intelligently generating a bathroom decoration plan. Background Art

[0002] With the acceleration of urbanization, the prosperity of the real estate industry has indeed led to the rise of the housing decoration market. After purchasing a new house, one of the first problems consumers face is how to decorate the interior to make it both beautiful and practical. However, the traditional bathroom design process has some limitations and challenges:

[0003] Personalized needs are difficult to meet: Each consumer's taste and needs are unique, and traditional design processes may find it difficult to fully meet personalized decoration styles and functional needs.

[0004] Differences in designer capabilities: Differences in the professional levels and aesthetic concepts of designers may result in design solutions that do not meet consumer expectations.

[0005] High communication costs: Communication between designers and consumers may require multiple round trips, which increases communication costs and time.

[0006] Low design efficiency: The process of manually building renderings is cumbersome and inefficient, and cannot quickly respond to changes in market and consumer demand.

[0007] Difficulty in cost control: In the traditional design process, material selection and cost control often rely on the designer’s experience, making it difficult to achieve an accurate budget.

[0008] In summary, the current formulation of interior decoration facility plans still needs to be improved and developed. Summary of the invention

[0009] The purpose of the present invention is to solve the defects of the above-mentioned prior art and provide a method and device for intelligently generating bathroom decoration schemes.

[0010] The present invention provides a method for intelligently generating a bathroom decoration scheme, comprising:

[0011] Step 1: Obtain the customer's decoration requirements and the actual photos of the interior space to be customized;

[0012] Step 2: Input the decoration requirements into the trained bathroom design solution generation model, and output the bathroom design solution through the model;

[0013] Step 3: Generate a rendering of the interior space to be customized using the bathroom design plan and the actual photos of the interior space to be customized.

[0014] According to the intelligent generation method of bathroom decoration scheme provided by the present invention, the bathroom design scheme generation model includes: a first sub-model and a second sub-model;

[0015] The first sub-model is used to select bathroom design solutions from the bathroom product library according to the customer's decoration requirements;

[0016] The second sub-model is used to convert the bathroom design solution generated by the first sub-model into a text prompt.

[0017] According to the intelligent generation method of bathroom decoration plan provided by the present invention, the text prompt at least includes: the budget and bathroom function corresponding to the bathroom design plan.

[0018] According to the intelligent generation method of bathroom decoration scheme provided by the present invention, the bathroom product material library includes: a bathroom scheme library and a bathroom material library;

[0019] The first sub-model is used to select bathroom design solutions from the bathroom product library according to the customer's decoration requirements, including:

[0020] According to the customer's decoration needs, the original plan is matched from the bathroom plan library first; if no match can be found or the customer is not satisfied, the bathroom design plan is determined based on the bathroom material library.

[0021] According to the intelligent generation method of bathroom decoration scheme provided by the present invention, the determining of bathroom design scheme based on the bathroom material library includes:

[0022] The trained material model is used to select corresponding materials in the bathroom material library, and corresponding products are combined based on the selected materials; and a bathroom design plan is generated based on all the combined products.

[0023] According to the intelligent generation method of bathroom decoration scheme provided by the present invention, the step 1 of obtaining the decoration requirements of the customer includes:

[0024] Collecting decoration customization information from customers; the decoration customization information includes: audio data and text data;

[0025] Converting the audio data into audio text data, and merging the audio text data with the text data to obtain integrated text data;

[0026] Performing data cleaning on the integrated text data to obtain cleaned text data;

[0027] Extracting keywords from the cleaned text data using a trained keyword extraction model;

[0028] All extracted keywords are analyzed to determine the decoration needs.

[0029] According to the intelligent generation method of bathroom decoration scheme provided by the present invention, the analysis of all extracted keywords to determine the decoration requirements includes:

[0030] Analyze the extracted keywords one by one, classify them according to themes, and determine the customer's decoration needs based on the classification results.

[0031] According to the intelligent generation method of bathroom decoration scheme provided by the present invention, the keyword extraction model is constructed based on the keyword library;

[0032] The keyword library is a keyword library that is continuously updated according to market trends and customer needs.

[0033] According to the intelligent generation method of bathroom decoration scheme provided by the present invention, the step 2 comprises:

[0034] Prioritize the renovation needs according to the emphasis or mention frequency of the keywords;

[0035] The priority sorting results are input into the trained bathroom design solution generation model to generate bathroom design solutions.

[0036] The present invention also provides a device for executing the above-mentioned bathroom decoration scheme intelligent generation method, comprising:

[0037] The acquisition unit is used to obtain the customer's decoration requirements and the actual photos of the interior space to be customized;

[0038] A prediction unit, used to input the decoration requirements into a trained bathroom design solution generation model, and predict the bathroom design solution through the model;

[0039] A generating unit is used to generate a rendering of the interior space to be customized by using the bathroom design scheme and the real shot of the interior space to be customized.

[0040] Beneficial effects:

[0041] The bathroom decoration plan intelligent generation method and device provided by the present invention input the customer's decoration needs into a trained bathroom design plan generation model, predict the customer's bathroom design plan through the model, and obtain the rendering of the interior space to be customized based on the customer's bathroom design plan and the actual photos of the interior space to be customized. This method not only improves the design efficiency and quality, but also provides customers with a more personalized and targeted design experience; it also saves costs for enterprises and improves market competitiveness. By automatically extracting keywords in communication content and quickly determining customer needs and budgets based on them, the time for demand analysis and budget evaluation in the traditional manual customization process is greatly shortened; it can also more accurately grasp the customer's needs and avoid design solutions that do not meet customer expectations due to misunderstandings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A flow chart of the intelligent generation method of bathroom decoration scheme provided by the present invention;

[0043] Figure 2 This is a structural block diagram of the intelligent generation device for bathroom decoration scheme provided by the present invention;

[0044] Figure 3 This is a schematic structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention is described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] The scheme provided by the present invention is suitable for generating interior decoration design schemes, and is particularly suitable for generating interior decoration design schemes for bathrooms.

[0047] Figure 1 The flow chart of the intelligent generation method of bathroom decoration scheme provided by the present invention is as follows: Figure 1 Said method comprises:

[0048] Step 1: Obtain the customer's decoration requirements and the actual photos of the interior space to be customized;

[0049] Step 2: Input the decoration requirements into the trained bathroom design solution generation model, and use the model to predict the bathroom design solution.

[0050] Step 3: Use the bathroom design plan and actual photos of the interior space to be customized to generate a rendering of the interior space to be customized.

[0051] Specifically, the actual photos of the interior space to be customized may also include a floor plan. The customer's decoration requirements may include at least: style information, size information, and budget information. Among them, the style information can be specifically: European, American, Chinese, modern and other styles. The embodiment of the style is mainly reflected in the door panels of the bathroom cabinet; the size information mainly includes the size of the bathroom cabinet. The budget information refers to all the expenses required for the overall decoration of the shower room. The bathroom design scheme generation model provided in this application has the following inputs: style information, size information, and budget information, and the output is a bathroom design scheme. Since the output bathroom design scheme is not intuitive enough, this application also generates an interior space rendering by combining the bathroom design scheme with the actual photos of the interior space to be customized. Through the interior space rendering, customers can intuitively and clearly see the specific details of the bathroom decoration scheme, thereby effectively reducing the communication cost between designers and consumers.

[0052] It is understandable that the execution subject of the present invention may be a bathroom intelligent customization device, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0053] The intelligent generation method of bathroom decoration scheme provided by the embodiment of the present invention inputs the customer's decoration needs into a trained bathroom design scheme generation model, predicts the customer's bathroom design scheme through the model, and obtains the rendering of the interior space to be customized based on the customer's bathroom design scheme and the actual photos of the interior space to be customized. This method not only improves the design efficiency and quality, but also provides customers with a more personalized and targeted design experience; it also saves costs for enterprises and improves market competitiveness. By automatically extracting keywords from communication content and quickly determining customer needs and budgets based on them, the time for demand analysis and budget evaluation in the traditional manual customization process is greatly shortened; it can also more accurately grasp the customer's needs and avoid design schemes that do not meet customer expectations due to misunderstandings.

[0054] Furthermore, the bathroom design solution generation model includes: a first sub-model and a second sub-model; the first sub-model is used to screen out bathroom design solutions from the bathroom product library according to the customer's decoration requirements; the second sub-model is used to convert the bathroom design solutions generated by the first sub-model into text prompts.

[0055] Specifically, the input of the first sub-model is style information, size information, and budget information, and the output is a bathroom design plan. In actual applications, customers input their decoration requirements into the first sub-model, and the first sub-model can automatically identify the style information, size information, and budget information included in the customer's decoration requirements, and then select bathroom design plans that meet customer needs from the bathroom product material library based on these style information, size information, and budget information. At the same time, since customers cannot directly obtain the quotation corresponding to the bathroom design plan and the function of the cabinet in the bathroom plan from the bathroom design plan, the bathroom design plan generated by the first sub-model is also converted into a text prompt through the second sub-model. The text prompt includes at least the following information: the budget corresponding to the bathroom design plan, the bathroom function, and may also include the length of the main cabinet, the length of the mirror cabinet, the shower room, etc.

[0056] The method provided by the present invention, on the one hand, generates a bathroom design plan through the first sub-model, avoiding the designer to extract the communication information with the customer and design the plan by himself, so that the efficiency of generating the bathroom design plan is higher; on the other hand, the first sub-model can accurately and completely capture the customer's decoration demand information, avoiding the phenomenon that the designer may miss a key point in the process of obtaining the decoration demand information, so that the generated bathroom design plan is more in line with the customer's needs; on the other hand, because the customer's decoration needs can be completely obtained through the first sub-model, the bathroom design plan can be output once according to the customer's decoration needs through the first sub-model, avoiding the defect of repeated communication between the designer and the customer, thereby reducing the communication cost between the designer and the customer. On the other hand, the present invention converts the generated bathroom design plan into a text prompt through the second sub-model, avoiding the process of the designer explaining the decoration plan to the customer, so that the consumer can immediately know the specific information of the design plan, thereby further reducing the communication cost between the designer and the customer.

[0057] Furthermore, the text prompt includes at least: the budget corresponding to the bathroom design plan, bathroom functions, and may also include: size (cabinet, mirror cabinet), whether there is a shower room, and decoration style.

[0058] The method provided by the present invention converts the bathroom design plan generated by the first sub-model into a text prompt through the second sub-model, indicating the budget, bathroom functional dimensions (cabinet, mirror cabinet), whether there is a shower room, and decoration style corresponding to the bathroom design plan, so that consumers can immediately know the specific information of the design plan, thereby further reducing the communication cost between designers and customers.

[0059] Furthermore, the bathroom product material library includes: a bathroom plan library and a bathroom material library; the first sub-model is used to screen out bathroom design plans from the bathroom product material library according to the customer's decoration needs, including: according to the customer's decoration needs, preferentially matching the original plans from the bathroom plan library; if no match can be found or the customer is not satisfied, determining the bathroom design plan based on the bathroom material library.

[0060] Specifically, the bathroom solution library is a solution library composed of pre-designed bathroom design solutions. The bathroom material library includes: components that constitute various bathroom products; among them, bathroom products can at least include: bathroom cabinets, shower rooms, toilets, faucets, and showers. Among them, the components corresponding to the bathroom cabinets include at least: cabinet bodies, door panels, handles, paints, sockets, lamps, hair dryer storage racks, countertops, and basins. The components corresponding to the shower room include at least: handles, glass, hinges, rails, and bases. The components corresponding to the toilet include at least: toilet body and toilet cover. The components corresponding to the faucet include at least: handles and body. The components corresponding to the shower include at least: body, handles, overhead spray, and water supply pipe.

[0061] In order to quickly select bathroom design solutions that meet customer needs, the present invention preferentially selects existing solutions from the bathroom solution library. Only when the customer's decoration needs do not match the existing solutions, it is necessary to select corresponding materials from the bathroom material library to generate a new bathroom design solution that meets the customer's decoration needs. In addition, the newly generated bathroom design solution can also be added to the bathroom solution library, so that the bathroom solution library can be updated according to the decoration needs of different customers.

[0062] The method provided by the present invention effectively improves the efficiency of generating bathroom solutions by preferentially matching the original solutions from the bathroom solution library; when all the solutions in the bathroom solution library do not match the customer's needs or the customer is not satisfied, a bathroom design solution that meets the customer's needs is determined based on the bathroom material library, so that the generated solution better meets the customer's needs.

[0063] Furthermore, the following is a detailed introduction on how to determine the bathroom design plan based on the bathroom material library, which specifically includes the following solutions:

[0064] Using the trained material model, select the corresponding materials in the bathroom material library, and combine them into corresponding products based on the selected materials; generate bathroom design plans based on all the combined products.

[0065] Specifically, the first sub-model either finds an existing solution from the bathroom solution library, or first composes a single product from the bathroom material library of each component product, and then combines the products. For example, if a new bathroom cabinet needs to be reassembled, the components are selected from "cabinet body, door panel, handle, paint, socket, lamp, hair dryer storage rack, countertop, basin", and then a new bathroom cabinet product solution is generated based on the selected components; the new shower room solution is selected from the component materials that compose the shower room: "handle, glass, hinge, track, base"; other toilets, faucets, showers, etc. are similar and will not be repeated in this invention.

[0066] The construction of the material model includes:

[0067] First, collect all the component information contained in the bathroom product library. Each component includes: the code corresponding to the component, and the component name, style, size, color, budget, material, function, etc. corresponding to the code. Secondly, pre-process the component information to obtain the processed component information; that is, organize all the component information into structured data format information, such as CSV files or database tables, for subsequent processing; then clean and standardize the data format information to obtain cleaned data, that is, processed component information; then, extract component features from each component information, and the component features are used to match customer needs; use the materials in the bathroom material library to train the initial material model to obtain a trained material model, the input of which is: decoration requirements, and the output is the component code.

[0068] The method provided by the present invention, on the one hand, uses a trained material model to screen corresponding materials in a bathroom material library, effectively improving the screening efficiency of materials, thereby improving the combination efficiency of using materials to combine into products, and ultimately improving the generation efficiency of bathroom design solutions; on the other hand, firstly, by screening materials that meet customer needs in the bathroom material library and combining them into corresponding products, the combined products are made to better meet the customer's decoration needs; then, based on the combined corresponding products, a bathroom design solution is finally generated, so that the finally generated bathroom design solution is more in line with the customer's decoration needs.

[0069] The following is an introduction to how to use the trained material model to select the corresponding materials in the bathroom material library:

[0070] First, the customer's decoration requirements are input into the trained material model, and the similarity or matching degree between the customer's decoration requirements and each material (component) in the bathroom material library is calculated through the material model; finally, the products are sorted according to the similarity or matching degree, and a list of materials that meet the customer's requirements is screened out. Among them, the calculation of the similarity or matching degree between the customer's decoration requirements and each component in the product material library can adopt a heuristic algorithm (such as a greedy algorithm, a simulated annealing algorithm, etc.) or an optimization algorithm (such as a genetic algorithm, a particle swarm optimization algorithm, etc.).

[0071] The following is an introduction to the construction of the bathroom material library:

[0072] Collect pictures and models of various bathroom products; mark the attributes of each component in the pictures and models of various bathroom products; create an index based on the attributes of each component to obtain a bathroom product material library.

[0073] Specifically, first, from the collected pictures and models of various bathroom products, the properties of all components (at least including: cabinet, paint, door panel, handle, functional parts, lamp, countertop, basin, shower room, faucet, shower) in the pictures and models are marked (at least including: name, size, material, budget, style, color, function), and then all the components are assigned corresponding codes respectively. Finally, according to the corresponding code of each component and its corresponding attributes, an index is created to obtain a bathroom material library.

[0074] The present invention provides an indexing method for a bathroom material library: taking a cabinet body as an example, it can be subdivided into a main cabinet with double doors, a main cabinet with double drawers, and a main cabinet with double door mirrors according to the name; the main cabinet with double doors can be divided into: multi-layer board with melamine, multi-layer board with PET, rubber wood, and multi-layer board with plain board according to the corresponding material; the main cabinet with double doors can be assigned a corresponding code according to the corresponding material and size, for example: the main cabinet with double doors with a size of 600*500*480 and a material of multi-layer board with melamine has a price of 500 / piece, and the main cabinet with double doors is assigned a code of Z-M001; and the main cabinet with the same size and a material of multi-layer board with PET has a price of 600 / piece, and the main cabinet with double doors is assigned a code of Z-M002... Take door panels as an example. According to their styles, door panels can be divided into European door panels, American door panels, Chinese door panels, and modern door panels. Then, European door panels can be assigned the code M-O001 based on their size, color, material, and price. American door panels can be assigned the code M-M001 based on their size, color, material, and price. Chinese door panels can be assigned the code M-Z001 based on their size, color, material, and price. Take paint as an example: the corresponding code can be assigned according to the color, style, material, price, and budget of the paint. For example, for paint with a color of milky white, a style of modern simplicity, European, modern, or American, a material of multilayer board, rubber wood, and a price of 1,000 yuan / meter, the code is Y-0001.

[0075] The following is an introduction on how to construct the second sub-model:

[0076] Collect a dataset containing bathroom design plans, which include at least: images and / or CAD drawings. For each bathroom design plan, collect the corresponding text prompts, which should contain detailed information such as size, total budget, function, shower room, etc.; convert the bathroom design plan data into a format that the model can handle. If the data is an image, image preprocessing (such as scaling, normalization, etc.) is required to obtain the preprocessed image; use a deep learning model such as a convolutional neural network (CNN) to extract image features of the preprocessed image; use the image features and the corresponding text prompts to train through the Transformer model, select a suitable loss function to optimize the model during the training process, and finally obtain a trained second sub-model.

[0077] The following is a detailed introduction on how to obtain the decoration needs of customers. The solution specifically includes: first, collecting the decoration customization information of customers; the decoration customization information includes: audio data and text data; converting the audio data into audio text data, and merging the audio text data and text data to obtain integrated text data; performing data cleaning on the integrated text data to obtain cleaned text data; using the trained keyword extraction model to extract keywords from the cleaned text data; analyzing all the extracted keywords to determine style information, size information, and budget information.

[0078] Specifically, the customer's decoration customization information can be obtained through the following methods:

[0079] Method 1: Recording with a mobile phone or tablet computer, extracting keywords such as size, style, budget, etc. from the conversation between the on-site salesperson and the customer; Method 2: Reading the customer's decoration design drawings, or reference pictures or PPTs, or videos, extracting the design style, bathroom size, and cost analysis. Method 3: Input size, style, budget, material, and other customization requirements in text.

[0080] The embodiment of the present invention uses speech recognition technology (ASR) to convert audio data into audio text data, merges the audio text data and text data to obtain integrated text data; secondly, performs data cleaning on the integrated text data to obtain cleaned text data; then, extracts keywords from the cleaned text data, and the keywords at least include style information, size information, budget information and other words closely related to bathroom design. These decoration requirements will directly reflect the needs and expectations of customers.

[0081] It can be understood that the decoration customization information includes the communication content between the staff and the customer and the questionnaire information provided by the customer. The communication content may include audio data and text data, and the questionnaire information is generally text data. In this embodiment, the audio data is converted into audio text data using speech recognition technology, specifically including: preprocessing the audio data, including noise suppression, audio enhancement and audio segmentation; applying window functions and fast Fourier transform (FFT) and other methods to extract audio features from the preprocessed audio data; inputting the extracted audio features into the speech recognition model for recognition, and obtaining the recognition results output by the speech recognition model; normalizing the recognition results, such as adding punctuation marks, adjusting the case, etc., optimizing the recognition results according to the context information, improving the accuracy, and finally converting the recognized and processed speech into audio text data.

[0082] In this embodiment, when preprocessing the audio data, noise reduction technology is applied to reduce background noise to improve the clarity of the voice signal, and audio enhancement technology is used to improve the voice quality, such as gain adjustment, echo cancellation, etc.

[0083] In this embodiment, when preprocessing audio data, if the audio contains multiple speakers, audio segmentation is required to distinguish different voice streams.

[0084] In this embodiment, data cleaning is performed on the integrated text data, including removing words, sentences and symbols irrelevant to bathroom design, such as modal particles, repeated words, punctuation marks, etc., from the integrated text data.

[0085] Formatting: unify the format of text data, such as converting uppercase letters to lowercase letters, removing extra spaces and line breaks, etc., to facilitate subsequent processing.

[0086] The method provided by the present invention effectively expands the scope of keyword acquisition by respectively acquiring keywords that meet the customer's decoration needs from audio data and text data; and uses the extracted keywords to determine style information, size information, and budget information, so that the determination of style information, size information, and budget information is more comprehensive and accurate, thereby further improving the design quality and efficiency of bathroom design solutions.

[0087] In addition, the present invention uses a trained keyword extraction model to extract keywords from the cleaned text data; the keywords include: cabinet, door panel, handle, paint color, board pattern, socket, lamp, hair dryer storage rack, countertop, basin, storage space, makeup, fill light, cosmetics preservation, economy, affordability, acceptable price, U-shaped, gap in the column, modern simplicity, European style, modern, American style, new Chinese color, light luxury, color, material, size.

[0088] The method provided by the present invention effectively reduces the communication cost between designers and consumers by utilizing a trained keyword extraction model to extract keywords from cleaned text data.

[0089] Further, all the extracted keywords are analyzed to determine the style information, size information, and budget information. Specifically, the extracted keywords are analyzed one by one and classified according to the theme to determine the customer's functional requirements, space layout requirements, customer preferred style and materials, and budget. For example:

[0090] By analyzing the keywords in the functional demand category, such as storage space, makeup, fill light, and cosmetics preservation, we can identify customers' specific needs for bathroom functions.

[0091] Analyzing keywords in material and style categories, such as “solid wood,” “metal,” “glass,” etc., can determine the styles and materials preferred by customers.

[0092] By analyzing the keywords of the space layout demand category, such as "U-shaped" and "gaps at the columns", we can determine the customer's needs for space layout.

[0093] By analyzing budget category keywords such as specific price, "economical", "affordable", "acceptable price", etc., you can determine the customer's budget.

[0094] The method provided by the present invention classifies the extracted keywords according to themes and determines the customer's decoration needs based on the classification results, so that the final bathroom design solution is more in line with the customer's expectations and the design efficiency is also improved.

[0095] Among them, the keyword extraction model can be trained by the following method: collect historical keywords to build a keyword library; use the keyword library as training samples; divide the training samples into training sets and test sets; use the training set to train the pre-trained ChatGPT model, and use the test set for testing and tuning, and finally obtain a trained keyword extraction model.

[0096] Specifically, a keyword library containing bathroom design related words is established, and the keyword library is used to form training samples to train the keyword extraction model. In this way, when extracting keywords from the communication content, the cleaned text data can be matched with the keyword library to obtain keywords.

[0097] The embodiment of the present invention uses a trained keyword extraction model to extract keywords from the cleaned text data. The keyword extraction model uses the ChatGPT model. Since the ChatGPT model has powerful natural language processing and context understanding capabilities, it can analyze a large amount of text data and identify key information and concepts in the text based on the knowledge learned during its training process. Therefore, this feature can be fully utilized when extracting decoration needs. By allowing the model to analyze the context of the entire conversation, rather than just focusing on a single sentence or phrase, keywords that reflect the customer's overall decoration needs can be more accurately extracted.

[0098] The following is an example of a specific application scenario to illustrate the keyword extraction process of the present invention. The scenario is: Customer A goes to a bathroom store to buy bathroom products, and clerk B receives Customer A. Clerk B turns on a device (the device can be a mobile phone or tablet computer, and recording is turned on) to obtain the customer's decoration needs.

[0099] B: Hello, do you want to buy a toilet?

[0100] A: Yes, I would like to buy bathroom cabinets, shower heads and other bathroom products together.

[0101] B: How many bathrooms do you have in your home?

[0102] A: 2.

[0103] B: Let me recommend some products to you regarding the total budget for two bathrooms.

[0104] A: The budget is relatively small, so I’d try to find one with the best value for money, probably around 40,000!

[0105] B: What is the decoration style of your home?

[0106] A: Modern and simple.

[0107] B: Then bathroom products should also be modern in style, right?

[0108] A: Yes, something like cream style, not too complicated, try to keep it simple.

[0109] B: How big can a bathroom cabinet be?

[0110] A: Both are probably 1 meter, right?

[0111] B: Let me recommend bathroom cabinets to you first, because bathroom cabinets are large products and can most affect the style of your bathroom.

[0112] A: OK.

[0113] B: So the budget for a bathroom cabinet is about 5,000 yuan, right? The toilet is 4,000, the shower head is 1,000, the shower room is 8,000, the towel rack is 500, the faucet is 500, and the rest is 1,000?

[0114] A. Yes, you can. The price of a bathroom cabinet is about 5,000 yuan.

[0115] B: What functions do you want the bathroom cabinet to have?

[0116] A: I need a bigger and deeper basin, because when I wash my hair now, water often splashes out and there is water everywhere.

[0117] B: Then what about a big square basin?

[0118] A: Okay, the bathroom I'm using now is quite dark, so I can add some lights to the mirror cabinet to provide some light. It would be best if there were sockets inside, as I'm used to blow-drying my hair in front of the bathroom cabinet.

[0119] B: We have a socket dedicated to the mirror cabinet, as you can see (showing the photo to A), and there is also a hair dryer storage rack that can be matched with the socket, so that the hair dryer can be placed next to the socket and there is no need to unplug the cord every time.

[0120] A: That's great. I have a lot of cosmetics. Is there any good storage solution?

[0121] B: We happen to have a mirror cabinet door storage rack that can store facial masks, lipsticks, paintbrushes, etc. (display storage rack).

[0122] A: This is good!

[0123] B: Then I can help you arrange the design and quotation?

[0124] A: OK.

[0125] Finally, the keywords extracted by the keyword extraction model include: "cost-effective", "modern style", "cream style, not too complicated, try to be simple", "1 meter", "budget 5000", "function", "bigger basin, deeper", "square large basin", "mirror cabinet", "light", "socket", "hair dryer storage rack", "storage", "mirror cabinet door panel storage rack".

[0126] Following the above example, the second sub-model provided by the embodiment of the present invention is further described by example:

[0127] The first sub-model matches the original solution from the bathroom solution library according to the above keywords; when the bathroom solution library cannot match the corresponding solution or the customer is not satisfied, the trained material model is used to screen the corresponding products in the bathroom product library. The product screening process is as follows:

[0128] 1. Main cabinet: Z-M001 (multi-layer board with melamine and wood cloth grain) + Z-C001 (400 long multi-layer board with melamine and wood cloth grain) = 600 + 400 = 1000, matching with "1 meter", modern and simple matching style "modern and simple", "high cost performance" matching with multi-layer board with melamine.

[0129] 2. Countertop: The “modern simplicity” in the keyword T-Y001 matches the “modern simplicity” in the material information library “modern simplicity, European style, modern, American style”, and the “2,000 yuan / m” and “high cost performance” are matched.

[0130] 3. Basin: T-P002 (matching key words: "large square basin").

[0131] 4. Mirror cabinet: Z-J002-matching main cabinet 1 meter, "3000 yuan" matching with high cost performance.

[0132] 5. Other configurations of the mirror cabinet: S-0001 (matching keywords "hair dryer storage rack, modern simple style"), S-0002 (matching: cream color, socket), S-0002 (matching: storage rack, facial mask storage, etc.).

[0133] Finally, the second sub-model is used to convert the generated bathroom design plan into a text prompt, which is: total budget: 38,000 yuan / set, main cabinet length: 1,000 mm, mirror cabinet length: 900 mm; functions: socket, hair dryer storage, cosmetics storage rack.

[0134] Furthermore, the keyword extraction model is constructed based on a keyword library; and the keyword library is a keyword library that is continuously updated according to market trends and customer needs.

[0135] Specifically, the establishment of the keyword library also includes: continuously updating the keyword library according to market trends and customer needs, and using the updated keywords to train the keyword extraction model.

[0136] The method provided by the present invention continuously updates the keyword library by utilizing market trends and customer needs, and utilizes the updated keywords to train the keyword extraction model, so that the updated keyword extraction model can more accurately extract customer needs, thereby making the ultimately generated bathroom design solution more in line with customer decoration needs.

[0137] Furthermore, the decoration requirements are input into the trained bathroom design solution generation model. Through the model, the bathroom design solution prediction also includes: prioritizing style information, size information, and budget information according to the degree or frequency of keyword mentions by customers; and inputting the priority sorting results into the trained bathroom design solution generation model to generate a bathroom design solution.

[0138] Specifically, in this embodiment, the demands can also be prioritized according to the degree of emphasis or frequency of mention of the keywords mentioned by the customers, and then the sorted customer demands can be input into the bathroom design solution generation model. Through the bathroom design solution generation model, a bathroom design solution that better meets the customer needs can be generated.

[0139] The intelligent generation method of bathroom decoration scheme provided by the present invention can more accurately grasp the real needs and budget expectations of customers and further reduce the possibility of misunderstanding and deviation by prioritizing keywords according to the degree of emphasis or frequency of mention and generating bathroom design schemes according to the sorting results.

[0140] The following is a detailed introduction on how to generate the interior space rendering to be customized, including the following solutions:

[0141] Preprocessing the real shot of the interior space to be customized; extracting structural information of the preprocessed real shot of the interior space to be customized; the structural information includes normal, depth and edge information of the interior space to be customized; generating an interior space image according to the bathroom product material library, the bathroom design plan, and the structural information; adjusting the interior space image to obtain a rendering of the interior space to be customized.

[0142] Specifically, a real shot of the interior space to be customized is obtained, and the real shot of the interior space to be customized is preprocessed; the preprocessed real shot of the interior space to be customized is input into the ControlNet model to extract the structural information of the interior space to be customized, and the structural information includes the normal, depth and edge information of the interior space to be customized; the Stable Diffusion model is combined with the 3D modeling software to generate an interior space image according to the bathroom product material library, the bathroom design plan, and the structural information of the interior space to be customized; the LoRA model is used to fine-tune the interior space image to obtain a rendering of the interior space to be customized.

[0143] The method provided by the present invention greatly reduces the time and labor intensity of manual operation by automatically extracting structural information and generating images using ControlNet and Stable Diffusion models. Moreover, through the preprocessing of real photos and the accurate extraction of structural information, a high degree of matching between the design scheme and the actual indoor space is ensured.

[0144] Furthermore, preprocessing of the actual photos of the customized interior space includes the following aspects:

[0145] Color Correction: Adjust the brightness, contrast, and color balance of an image to ensure accurate image colors.

[0146] Resizing: Adjust the resolution and size of the image according to the input requirements of the ControlNet model.

[0147] Image Cropping: Remove unnecessary background or edges from an image to make the interior space the main focus.

[0148] Format Conversion: Convert images to a format supported by the ControlNet model, such as PNG or JPEG.

[0149] The method provided by the present invention can estimate the normal direction of each pixel by analyzing the pixels and textures in the real-life picture of the indoor space to be customized through the ControlNet model. Through the deep learning algorithm, ControlNet can infer the relative depth of objects in the image, that is, their distance from the camera. ControlNet can identify the edges of objects by analyzing changes in color, brightness, etc. in the image. The normal, depth and edge information output by the ControlNet model can also be post-processed, such as smoothing, denoising, etc., to improve the accuracy and availability of the information. In addition, the embodiment of the present invention uses the structural information of the indoor space to be customized output by the ControlNet model as a constraint condition to guide the Stable Diffusion model to follow these structures when generating images, while retaining the creative output of the Stable Diffusion model in terms of texture, lighting, color, etc.

[0150] Furthermore, the embodiments of the present invention are combined with 3D modeling software to place bathroom products in the design scheme in the indoor space in 3D form, and use the Stable Diffusion model for rendering to generate realistic indoor space images. Specifically, the following steps are used: according to the structural information of the actual photos, the basic framework of the indoor space is established in the 3D modeling software; the elements in the design scheme are converted into 3D models, and the models in the material library are used for layout and placement; and the Stable Diffusion model is used to generate indoor space images according to the bathroom design scheme and the 3D model.

[0151] The 3D modeling software may be 3D Studio Max (3DS MAX), Maya, SketchUp, etc. The LoRA model allows for fine-tuning of certain parts of the model while maintaining the stability of the original model to achieve a specific style or feature. By loading the weights of the LoRA model, the indoor space image generated by the Stable Diffusion model can be fine-tuned to better meet the requirements of the design plan.

[0152] The method provided by the present invention generates a rendering by combining the bathroom design plan with the real photos of the interior space to be customized, so that customers can intuitively see the effect of the design plan in the actual space. This visual presentation method helps customers better understand the design plan, improve decision-making efficiency, and reduce rework caused by misunderstanding or imagination deviation.

[0153] Furthermore, the present invention can also collect historical bathroom design plans and corresponding historical interior space real photos and historical renderings to train the Stable Diffusion model, ControlNet model and LoRA model.

[0154] The intelligent generation device for bathroom decoration scheme provided by the present invention is described below. The intelligent generation device for bathroom decoration scheme described below and the intelligent generation method for bathroom decoration scheme described above can be referred to each other.

[0155] Figure 2 The structure diagram of the intelligent generation device for bathroom decoration scheme provided by the present invention is as follows: Figure 2 As shown, the device comprises:

[0156] The acquisition unit 201 is used to acquire the customer's decoration requirements and the real photos of the interior space to be customized; the decoration requirements include: style information, size information, and budget information;

[0157] Prediction unit 202, used to input the decoration requirements into the trained bathroom design solution generation model, and predict the bathroom design solution through the model;

[0158] The generating unit 203 is used to generate a rendering of the interior space to be customized by using the bathroom design scheme and the real shot of the interior space to be customized.

[0159] Figure 1 shows a schematic diagram of the physical structure of an electronic device. Figure 3 As shown, the electronic device may include: a processor 310, a communications interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communications interface 320 and the memory 330 communicate with each other through the communication bus 340. The processor 310 may call the logic instructions in the memory 330 to execute the intelligent generation method of the bathroom decoration plan, which includes:

[0160] Step 1: Obtain the customer's decoration requirements and real photos of the interior space to be customized; the decoration requirements include: style information, size information, and budget information;

[0161] Step 2: Input the decoration requirements into the trained bathroom design solution generation model, and use the model to predict the bathroom design solution;

[0162] Step 3: Generate a rendering of the interior space to be customized using the bathroom design plan and the actual photos of the interior space to be customized.

[0163] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0164] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the bathroom decoration scheme intelligent generation method provided by the above methods, the method includes:

[0165] Step 1: Obtain the customer's decoration requirements and the actual photos of the interior space to be customized; decoration requirements include: style information, size information, and budget information;

[0166] Step 2: Input the decoration requirements into the trained bathroom design solution generation model, and use the model to predict the bathroom design solution;

[0167] Step 3: Use the bathroom design plan and actual photos of the interior space to be customized to generate a rendering of the interior space to be customized.

[0168] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the bathroom decoration scheme intelligent generation method provided by the above methods, the method comprising:

[0169] Step 1: Obtain the customer's decoration requirements and the actual photos of the interior space to be customized; decoration requirements include: style information, size information, and budget information;

[0170] Step 2: Input the decoration requirements into the trained bathroom design solution generation model, and use the model to predict the bathroom design solution;

[0171] Step 3: Use the bathroom design plan and actual photos of the interior space to be customized to generate a rendering of the interior space to be customized.

[0172] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0173] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligently generating bathroom decoration schemes, characterized in that: include: Step 1: Obtain the customer's decoration requirements and the actual photos of the interior space to be customized; Step 2: Input the decoration requirements into the trained bathroom design solution generation model, and output the bathroom design solution through the model; Step 3: Generate a rendering of the interior space to be customized using the bathroom design plan and the actual photos of the interior space to be customized.

2. The intelligent generation method of bathroom decoration scheme according to claim 1 is characterized in that: The bathroom design solution generation model includes: a first sub-model and a second sub-model; The first sub-model is used to select bathroom design solutions from the bathroom product library according to the customer's decoration requirements; The second sub-model is used to convert the bathroom design solution generated by the first sub-model into a text prompt.

3. The intelligent generation method of bathroom decoration scheme according to claim 2 is characterized in that: The text prompt at least includes: the budget and bathroom functions corresponding to the bathroom design plan.

4. The intelligent generation method of bathroom decoration scheme according to claim 2 is characterized in that: The bathroom product material library includes: a bathroom solution library and a bathroom material library; The first sub-model is used to select bathroom design solutions from the bathroom product library according to the customer's decoration requirements, including: According to the customer's decoration needs, the original plan is matched from the bathroom plan library first; if no match can be found or the customer is not satisfied, the bathroom design plan is determined based on the bathroom material library.

5. The intelligent generation method of bathroom decoration scheme according to claim 4 is characterized in that: Determining the bathroom design solution based on the bathroom material library includes: The trained material model is used to select corresponding materials in the bathroom material library, and corresponding products are combined based on the selected materials; and a bathroom design plan is generated based on all the combined products.

6. The intelligent generation method of bathroom decoration scheme according to any one of claims 1 to 5, characterized in that: The decoration requirements of customers mentioned in step 1 include: Collecting decoration customization information from customers; the decoration customization information includes: audio data and text data; Converting the audio data into audio text data, and merging the audio text data with the text data to obtain integrated text data; Performing data cleaning on the integrated text data to obtain cleaned text data; Extracting keywords from the cleaned text data using a trained keyword extraction model; All extracted keywords are analyzed to determine the decoration needs.

7. The intelligent generation method of bathroom decoration scheme according to claim 6 is characterized in that: The analysis of all extracted keywords to determine the decoration requirements includes: Analyze the extracted keywords one by one, classify them according to themes, and determine the customer's decoration needs based on the classification results.

8. The intelligent generation method of bathroom decoration scheme according to claim 6 is characterized in that: The keyword extraction model is constructed based on the keyword library; The keyword library is a keyword library that is continuously updated according to market trends and customer needs.

9. The intelligent generation method of bathroom decoration scheme according to claim 6 is characterized in that: The second step comprises: Prioritize the renovation needs according to the emphasis or mention frequency of the keywords; The priority sorting results are input into the trained bathroom design solution generation model to generate bathroom design solutions.

10. A device for executing the bathroom decoration plan intelligent generation method according to claim 1, characterized in that: include: The acquisition unit is used to obtain the customer's decoration requirements and the actual photos of the interior space to be customized; A prediction unit, used to input the decoration requirements into a trained bathroom design solution generation model, and predict the bathroom design solution through the model; A generating unit is used to generate a rendering of the interior space to be customized by using the bathroom design scheme and the real shot of the interior space to be customized.

Images