I-Design: Personalized LLM Interior Designer

Ata Çelen,Guo Han,Konrad Schindler,Luc Van Gool,Iro Armeni,Anton Obukhov,Xi Wang

2024-04-04

Abstract:Interior design allows us to be who we are and live how we want - each design is as unique as our distinct personality. However, it is not trivial for non-professionals to express and materialize this since it requires aligning functional and visual expectations with the constraints of physical space; this renders interior design a luxury. To make it more accessible, we present I-Design, a personalized interior designer that allows users to generate and visualize their design goals through natural language communication. I-Design starts with a team of large language model agents that engage in dialogues and logical reasoning with one another, transforming textual user input into feasible scene graph designs with relative object relationships. Subsequently, an effective placement algorithm determines optimal locations for each object within the scene. The final design is then constructed in 3D by retrieving and integrating assets from an existing object database. Additionally, we propose a new evaluation protocol that utilizes a vision-language model and complements the design pipeline. Extensive quantitative and qualitative experiments show that I-Design outperforms existing methods in delivering high-quality 3D design solutions and aligning with abstract concepts that match user input, showcasing its advantages across detailed 3D arrangement and conceptual fidelity.

Artificial Intelligence

What problem does this paper attempt to address?

The problem addressed by this paper is how to make interior design more personalized and user-friendly to meet users' design needs and preferences.

I-Design is a personalized large language model (LLM) interior designer that allows users to describe their design goals using natural language and generate feasible 3D scene graphs. The system first transforms the text input from users into scene graphs through dialogues and logical reasoning by a set of LLM agents, and then uses a backtracking algorithm to determine the optimal positions of each object in the scene. Finally, it retrieves and integrates assets from an existing 3D object database to build the final 3D design. Additionally, the paper proposes a new evaluation protocol that utilizes a visual-language model to complement the design process.

The research indicates that existing 3D interior scene synthesis methods have limitations in understanding and implementing users' abstract concepts. I-Design overcomes this issue through collaborative LLM agents that can comprehend design principles, styles, and spatial layouts. Experiments show that I-Design outperforms existing methods in generating high-quality 3D design solutions and aligning them with users' input concepts.