What visual language can emerge when choreography is read as drawing in space? "F O R M S" is a digital art concept that brings together dance and machine learning—specifically real-time human pose detection—to compose a live, interactive visual experience. The work asks how technology can deepen, rather than replace, the poetics of the moving body.
The project explores the relationship between dance and visual art by generating abstract and literal visual models from the dancer's movement. Its aim is to enhance the perception of choreography: to add a new layer of visual composition that runs parallel to gesture, rhythm, and presence. A state-of-the-art pose-detection pipeline analyses the positions and trajectories of the body in real time, so that each phrase of movement can be translated into a distinct visual form as it happens.
Technically, the framework builds on human pose estimation—tracing keypoints and their relations in space—to drive responsive graphics. Early studies sketched how connections between extremities, between distributed points, and across the whole figure could each suggest a different visual grammar: from spare lines between hands and feet to denser networks of relations and, finally, configurations that echo the human form itself. Those experiments became the basis for algorithmic components that render live graphics from detected poses.
"F O R M S" can unfold as installation or live performance, inviting performers and public to inhabit the same responsive field. The ambition is twofold: to offer audiences a fresh visual reading of dance, and to give dancers a mirror for their own material—expanding physical vocabulary, clarifying intention in the limbs, and supporting creativity through a new visual understanding of movement. The work also carries pedagogical potential, nurturing body awareness and a more conscious relationship to movement from an artistic and educational perspective.
A case study integrated "F O R M S" into a live performance with professional young dancers, who could choose among several visual models to align with their choreographic choices. The piece turned on how we see the world and one another—on perspective—and each form explored that idea in movement. After a period of rehearsal and real-time integration, participants reflected on the experience in overwhelmingly positive terms: many described clearer attention to range and precision, a sense of seeing themselves anew, and a desire for more expansive, expressive movement.
First developed in dialogue with the contemporary dance work "Perspectives" (June 2022), "F O R M S" reframed the stage as an adaptive environment where machine perception and choreography meet. The intersection of dance and learning-based systems here is not only technical but artistic: it proposes an interdisciplinary path between art and technology, and opens toward future work—larger ensembles, broader publics, and experiences that extend well-being and embodied awareness through interaction.
In that sense, "F O R M S" contributes to a wider conversation on interactive art, neural perception, and human-centred design: machine learning as a means to amplify and extend the boundaries of artistic expression, while keeping the body—and its poetry—at the centre of the work.