Exploring Facial Recognition in Interactive Visuals with TouchDesigner: The Case of Expression Topology

This study explores the application of facial recognition technology in interactive visual art through the original work Expression Topology. By capturing real-time facial expression data from viewers, the system drives dynamic image generation, forming a “face-data-visual” interactive loop. Built with TouchDesigner, and integrated with OpenCV and MediaPipe, the system extracts precise facial feature data and maps it to visual variables through parametric design. The artwork uses abstract and topological imagery to translate emotional expressions into generative visuals, emphasizing the viewer’s role in shaping the narrative. Results show that this approach enhances responsiveness and visual diversity in interactive works. The unpredictability of expressions enriches the generative process, highlighting the value of real-time interaction based on biometric input in digital art.