AI Technology for Digital Media in Spatial Interactive Design of Immersive Art Experience Hall under AM-MF-ELM Algorithm

This study aims to address the problems of insufficient interaction between audiences and artworks, and the disconnection between technical application and artistic expression in traditional art exhibitions. It seeks to create a richer, more personalized, and immersive artistic experience for audiences. First, the key roles of digital media and Artificial Intelligence (AI) in immersive artistic experiences are sorted out. Digital media technology solves the problems of visual presentation and in-depth experience. This study innovatively integrates flow experience theory with immersive technology to explore the psychological mechanism through which technical parameters influence artistic experience, and constructs an original Technology-Art-Cognition (TAC) coupling model for immersive art space interaction design. On this basis, a spatial interaction framework centered on multimodal gesture recognition technology is designed, in which audiences can guide virtual light and shadow and trigger artistic scenes through natural gestures, realizing the deep coupling between artistic expression and audience intention. In addition, an improved Extreme Learning Machine based on Attention Mechanism and Multi-Feature Fusion (AM-MF-ELM) is proposed for gesture recognition. Through multimodal perception and adaptive interaction design, an immersive experience system with deep integration of technology and art is constructed. Experimental results show that the improved AM-MF-ELM algorithm maintains excellent performance in all evaluation indicators, achieving 98.6% recognition accuracy on the high-precision gesture dataset, a 1.4% improvement compared with the traditional ELM algorithm (from 97.2% to 98.6%) and outperforming the SVM classifier (92.5%) by 6.1%. These results demonstrate that the attention mechanism and the multi-feature fusion strategy effectively enhance feature extraction capability and noise robustness, thereby providing reliable technical support for natural interaction in immersive art spaces. This study successfully establishes a spatial interaction design paradigm driven by both digital media and AI and realizes the interaction between audiences and artworks through multimodal gesture recognition technology, improving the quality and personalization of immersive artistic experiences.