Emotional Response Mechanism of Smart Textiles Integrated with Transformer Model in Garden Interactive Art Installation 

Tianlong Chai , Chengcheng Sha

Article
2026 / Volume 9 / Pages 80-99
Received 30 July 2025; Accepted 27 October 2025; Published 20 January 2026
https://doi.org/10.31881/TLR.2026.080

Abstract
To address the challenge that smart textiles face in achieving multimodal emotion perception and personalized artistic responses in complex garden environments, this paper proposes an emotion recognition and artistic feedback mechanism based on a lightweight Transformer model. By integrating electromyography (EMG), galvanic skin response (GSR), and voice sensors within smart textiles, users’ physiological signals and environmental data are collected, and multi-source information is fused to extract emotional characteristics. Combined with a metaphorical emotion-art mapping strategy, the light-emitting diode (LED) array and shape memory materials in the textiles are driven to generate dynamic color and morphological changes. This mechanism is applied to interactive garden art installations that integrate natural elements such as plants and terrain, with the generated artistic feedback synergistically responding to the natural environment. The model is optimized through a lightweight design, enabling real-time deployment. Experimental results show that the system achieves an average EMG signal-to-noise ratio (SNR) of 28.3 dB in interactive garden art installations and compresses the response delay to 190 ms in complex garden scenes. User satisfaction surveys indicate that the scores for joy and calmness emotional feedback are 4.25 and 4.05, respectively. This paper provides a feasible technical path for the application of smart textiles in the field of emotional interaction and art installations.

Keywords
smart textiles, interactive garden art installations, transformer model, emotion perception, multimodal data fusion