An interactive indoor layout generation and management system

Abstract

The present application belongs to the field of interactive indoor layout, and particularly relates to an interactive indoor layout generation and management system, comprising an analyst intelligent agent, a designer intelligent agent, an evaluator intelligent agent and a human feedback module. The present application realizes a complete process from requirement understanding to design execution, to layout evaluation and finally to manual modification and confirmation by combining the analyst intelligent agent with strong natural language understanding capability, the designer intelligent agent with strong natural language understanding capability, the evaluator intelligent agent with strong natural language understanding capability and the interaction with artificial, so that the system can efficiently convert user requirements into specific indoor design scheme, and through the cooperation and feedback cycle among intelligent agents, the continuous optimization and improvement of the design scheme are ensured, and the problems occurring in the design process can be intelligently classified and evaluated, so that each step of design can meet the actual application requirements and aesthetic standards of users.

Description

Technical Field

[0001] This invention relates to the field of interactive interior layout technology, and more particularly to an interactive interior layout generation and management system. Background Technology

[0002] Driven by computer vision and natural language processing technologies, the automation of interior layout design has become a research hotspot, widely applied in fields such as smart homes, game interface design, and virtual reality. The core challenge in interior layout generation lies in how to automatically generate practical and aesthetically pleasing interior design schemes based on user needs and spatial conditions. Traditional methods typically rely on designers' experience for manual layout, which is not only time-consuming but also difficult to handle complex or personalized needs.

[0003] With the advancement of technology, especially the rapid development of artificial intelligence and deep learning algorithms, the field of interior layout design is undergoing a transformation. These advanced technologies enable design systems to more accurately understand complex design requirements and automatically generate interior layout schemes that meet user expectations. For example, a study on Diffuscene (Tang J, Nie Y, Markhasin L, et al. Diffuscene: Denoising diffusion models for generative indoor scene synthesis[C] / / Proceedings of the IEEE / CVF conference on computer vision and pattern recognition. 2024: 20507-20518.) demonstrates how to use deep learning technology to analyze simple descriptions provided by users to generate preliminary interior layout schemes. This system, by learning from a large amount of interior design data, is able to identify and apply common elements and layout rules in the design. Furthermore, a study titled Layoutgpt (Feng W, Zhu W, Fu T, et al. Layoutgpt: Compositional visual planning and generation with large language models[J]. Advances in Neural Information Processing Systems, 2024, 36.) introduces a system that uses natural language processing (NLP) technology to directly generate and adjust interior designs from users' verbal needs. By collecting user input through dialogue, the system uses deep learning models to parse and transform these needs into specific design actions, thereby automating the entire design process without human intervention. The application of these technologies not only improves the efficiency of interior design but also helps designers overcome the limitations of traditional design methods in terms of innovation and responsiveness through automatically generated solutions. Deep learning and NLP technologies demonstrate enormous potential in processing user input and transforming it into practical design solutions, particularly in understanding and implementing users' personalized needs. These advancements indicate that artificial intelligence is becoming an indispensable tool in the field of interior design, greatly driving technological innovation and development in this field. However, these existing technologies typically face a major challenge: they are mostly trained on fixed datasets, resulting in insufficient generalization when dealing with unseen interior space configurations. This is particularly true when dealing with non-standard apartment layouts, where the adaptability and effectiveness of existing systems are often unsatisfactory. This problem is prominently illustrated in Diffuscene's research; while their system can handle simple spaces and interactions, its performance significantly degrades when encountering atypical or highly flexible spatial layouts. Furthermore, Layoutgpt is also limited to simple spatial interactions and design, with relatively low accuracy and rationality in spatial layout. Therefore, we propose an interactive interior layout generation and management system. Summary of the Invention

[0004] This invention proposes an interactive indoor layout generation and management system to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: An interactive interior layout generation and management system includes an analyst agent, a designer agent, an evaluator agent, and a human feedback module. The analyst agent is used to enable analysts to gain a deep understanding of user needs and output design language that designers can understand and execute. The designer agent is used to enable designers to conduct preliminary interior layout design based on task guidance issued by analysts; The evaluator AI agent is used to evaluate the layout output by the user's needs and the designer AI agent, raise layout issues, and indicate which opinions need to be reviewed manually. The human feedback module is used to allow users to make judgments and modifications based on the opinions provided by the evaluator agent that require evaluation.

[0006] Preferably, the analyst agent can also understand the opinions and questions raised by humans and evaluators, and output new layout modification tasks.

[0007] Preferably, the designer agent can also analyze the latest layout modification opinions from the analyst agent and adjust the design scheme accordingly.

[0008] Preferably, the core of the analyst agent, designer agent, and evaluator agent is a large natural language understanding model.

[0009] Preferably, the user requirements are described by the analyst agent's natural language understanding model to describe the user's specific design needs, such as a European-style living room, a three-seater sofa, and a square dining table.

[0010] Preferably, the spatial layout generation is presented in a structured code language through a large model that understands the natural language of the designer's intelligent agent, describing in detail the object categories, attributes, and spatial relationships of each space, such as the size, material, and location coordinates of the sofa.

[0011] Preferably, the assessor agent uses its own natural speech understanding model to classify problems according to their severity and records the problem type and whether human intervention is required.

[0012] Preferably, the problem is mainly divided into type I layout problem, type II layout problem and type III layout problem.

[0013] Preferably, the Type I layout problem is the problem of severe clipping that affects visual perception, the Type II layout problem is the problem of unconventional spatial layout, and the Type III layout problem is the problem of unattractive layout.

[0014] Compared with existing technologies, this invention combines analyst agents with strong natural language understanding capabilities, designer agents with strong natural language understanding capabilities, and evaluator agents with strong natural language understanding capabilities, along with human interaction, to achieve a complete process from needs understanding to design execution, layout evaluation, and final human modification and confirmation. This enables the system to efficiently transform user needs into specific interior design solutions. At the same time, the collaboration and feedback loop between agents ensures the continuous optimization and improvement of the design solutions. It can also intelligently classify and evaluate problems that arise during the design process, ensuring that each step of the design meets the user's actual application needs and aesthetic standards. Attached Figure Description

[0015] Figure 1 This is a connection block diagram of an interactive indoor layout generation and management system proposed in this invention; Figure 2 This is a structural diagram of the analyst agent, designer agent, and evaluator agent in this invention. Detailed Implementation

[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0017] Reference Figure 1-2 An interactive interior layout generation and management system, including analyst agent, designer agent, evaluator agent and human feedback module; The analyst agent is used to enable analysts to gain a deep understanding of user needs and output design language that designers can understand and execute. The designer agent is used to enable designers to conduct preliminary interior layout designs based on task guidance issued by analysts; The evaluator AI agent is used to evaluate the layout output by the user's needs and the designer AI agent, raise layout issues, and indicate which opinions need to be reviewed manually. The human feedback module is used to allow users to make judgments and modifications based on the opinions given by the evaluator agent that require evaluation; The analyst agent can also understand the opinions and questions from human and evaluator feedback and output new layout modification tasks. The designer agent can also adjust the design scheme based on the latest layout modification opinions from the analyst agent.

[0018] More specifically, the core of the analyst agent, designer agent, and evaluator agent is a large-scale natural language understanding model. User needs are described through the analyst agent's large-scale natural language understanding model, specifying the user's design requirements, such as a European-style living room, a three-seater sofa, and a square dining table. The spatial layout is generated through the designer agent's large-scale natural language understanding model, presented in a structured code language, detailing the object categories, attributes, and spatial relationships of each space, such as the size, material, and location coordinates of the sofa. The evaluator agent uses its own large-scale natural language understanding model to classify problems according to their severity, recording the problem type and whether human intervention is required. Problems are mainly divided into Type I, Type II, and Type III layout problems. Type I layout problems are those where clipping severely affects visual perception; Type II layout problems are those where the spatial layout is not conventional; and Type III layout problems are those where the layout is not aesthetically pleasing.

[0019] It should be noted that this embodiment combines analyst agents with strong natural language understanding capabilities, designer agents with strong natural language understanding capabilities, and evaluator agents with strong natural language understanding capabilities, along with human interaction, to achieve a complete process from needs understanding to design execution, layout evaluation, and final human modification and confirmation. This enables the system to efficiently transform user needs into specific interior design solutions. At the same time, the collaboration and feedback loop between agents ensures the continuous optimization and improvement of the design solutions. It can also intelligently classify and evaluate problems that arise during the design process, ensuring that each step of the design meets the user's actual application needs and aesthetic standards.

[0020] The specific implementation steps of this embodiment are as follows: S1: The analyst agent enables analysts to gain a deep understanding of user needs and output design language that designers can understand and execute. S2: The designer agent enables designers to conduct preliminary interior layout design based on task guidance issued by analysts; S3: The evaluator AI evaluates the layout output by the user's needs and the designer AI, raises layout issues, and points out which opinions need to be reviewed manually. S4: The human feedback module allows users to make judgments and modifications based on the opinions given by the evaluator agent that need to be evaluated; S5: The analyst agent understands the opinions and questions raised by humans and reviewers, and outputs new layout modification tasks; S6: The designer agent re-analyzes the latest layout modification suggestions from the analyst agent and adjusts the design scheme accordingly; S7: Repeat steps S3-S6 until the evaluator agent no longer raises new opinions, and finally output the home layout result that meets the user's needs and aesthetics.

[0021] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An interactive indoor layout generation and management system, characterized in that, This includes analyst agents, designer agents, assessor agents, and a human feedback module; The analyst agent is used to enable analysts to gain a deep understanding of user needs and output design language that designers can understand and execute. The designer agent is used to enable designers to conduct preliminary interior layout design based on task guidance issued by analysts; The evaluator AI agent is used to evaluate the layout output by the user's needs and the designer AI agent, raise layout issues, and indicate which opinions need to be reviewed manually. The human feedback module is used to allow users to make judgments and modifications based on the opinions provided by the evaluator agent that require evaluation.

2. The interactive indoor layout generation and management system according to claim 1, characterized in that, The analyst agent can also understand the opinions and questions raised by humans and reviewers, and output new layout modification tasks.

3. The interactive indoor layout generation and management system according to claim 2, characterized in that, The designer agent can also analyze the latest layout modification suggestions from the analyst agent and adjust the design scheme accordingly.

4. The interactive indoor layout generation and management system according to claim 3, characterized in that, The core of the analyst agent, designer agent, and assessor agent is a large natural language understanding model.

5. The interactive indoor layout generation and management system according to claim 4, characterized in that, The user requirements are described using a large model of natural language understanding by the analyst agent to describe the user's specific design needs.

6. The interactive indoor layout generation and management system according to claim 5, characterized in that, The spatial layout generation is presented in a structured code language through the natural language understanding of the designer agent's large model, which describes in detail the object categories, attributes, and spatial relationships of each space.

7. The interactive indoor layout generation and management system according to claim 6, characterized in that, The assessor agent uses its own natural speech understanding model to classify problems according to their severity and records the problem type and whether human intervention is required.

8. The interactive indoor layout generation and management system according to claim 7, characterized in that, The problems are mainly divided into type I layout problems, type II layout problems, and type III layout problems.

9. An interactive indoor layout generation and management system according to claim 8, characterized in that, The first type of layout problem is that clipping severely affects visual perception; the second type of layout problem is that the spatial layout is not unconventional; and the third type of layout problem is that the layout is not aesthetically pleasing.

Images