OSW

SIGNATURE WORK
CONFERENCE & EXHIBITION 2023

EVALUATING THE GENERATED AFFECTIVE ENVIRONMENTS USING STABLE DIFFUSION FOR INTERIOR DESIGN IN VIRTUAL REALITY

Name

Xuening Peng

Major

Data Science

Class

2023

About

I am Xuening Peng from Class of 2023. I major in Data Science and have a strong interest in the research field of Human-Computer Interaction.

Signature Work Project Overview

This research project aims to address gaps in previous studies investigating the affective impact of interior environments by exploring the impact of interior objects in conjunction with the environment using a generative tool in a virtual reality setting. The study used Stable Diffusion to generate interior spaces while considering shapes, colors, and textures of objects to elicit affective responses. A mixed-method approach was used to evaluate the human-AI trust of the tool, with valence and arousal reported by participants. The research addresses two research questions: (1) what affective responses can be elicited from generative interior space, and (2) how do people perceive the generative AI tool in the application of interior space design? The results showed generally positive feedback from participants regarding the tool’s ability to generate 360 panoramic visualizations of their ideal living rooms, although limitations were also identified. This study contributes to the growing research on the emotional requirements for comfortable living and suggests the potential of generative tools for interior space design.

Signature Work Presentation Video