YCAM Dance Crew 2024

Development of an AI Mirror to Unlock New Possibilities in Dance Expression

2024

  • rectangle large type 2 744763586f2b135df8bb98366769bb2d.jpeg
  • ycam-02
  • ycam-03
  • SDIM0562.JPG

We were responsible for the engineering of the AI system and interactive technology for the "YCAM Dance Crew 2024: Dance Booth Exhibition" exhibition organized by the Yamaguchi Center for Arts and Media [YCAM].

In this exhibition, participants can input text prompts (instructions) to use an image generation AI that adds effects to dance videos or transforms the videos into entirely different visuals. For those who may find it difficult to create prompts, we have implemented a mode where sentences are automatically generated by simply combining suggested words, making the experience accessible and easy for everyone.

Furthermore, the participants' dance performances were recorded, and the footage was transformed using AI-based video processing. After the experience, a QR code was displayed, allowing participants to access and download their personalized video content at a later time.

We used TouchDesigner as the main tool and adopted Stream Diffusion, a technology adapted for real-time processing based on Stable Diffusion, as the image generation AI. For prompt generation, we utilized a large language model (LLM), enabling participants to create natural instruction text by combining suggested words. To transform participants’ dance videos, we used a combination of AnimateDiff and IP-Adapter, achieving high-quality video transformation.

YCAM-20workflow-20(1)

System diagram

This exhibition aims to provide dancers with opportunities to discover and learn new choreography and movements, while also offering an enjoyable experience for a wide range of participants—including children and those new to dance—to move their bodies and feel more connected to dance. The AI mirror, which alters visuals in response to body movements, inspires new choreographic ideas and creates a novel experience that expands the possibilities of physical expression.

- Tokyo Art Beat - 「YCAM Dance Crew 2024 メディア・テクノロジーでダンスをパワーアップするワークショップ&ダンスブース」 - Qosmo Lab - AIを使って新しいダンス表現を引き出す [YCAM Dance Crew 2024]

Exhibition

2024.06.29 - 2024.11.10 | YCAM Dance Crew 2024: Dance Booth Exhibition (Yamaguchi Center for Arts and Media [YCAM])

Credits

[YCAM Dance Crew 2024: Dance Booth Exhibition] Technical Direction, Sound Design: Junji Nakaue (YCAM) System Design/development: Junji Nakaue(YCAM), Ryosuke Nakajima (Qosmo), Keito Takaishi (Qosmo), Yohei Miura (YCAM) Hardware Design/Production: Hayato Itakura (YCAM), Hazuki Ohta (YCAM) UI Design: Ayumi Takahara (YCAM) Dance Music Production: toiret status, nulye, DJJJ Assistance in Editing Exhibition Captions: Daichi Yamaoka (YCAM), Haruki Ohba (YCAM), Keina Konno (YCAM) Produced and administration: Akiko Takeshita (YCAM), Kilarla Akiyama (YCAM)

Host: Yamaguchi City Foundation for Cultural Promotion Supported: Yamaguchi City, Yamaguchi City Board of Education Funded: Agency for Cultural Affairs, Government of Japan, Fiscal Year 2024|Japan Arts Council Technical Support: YCAM InterLab Production: Yamaguchi Center for Arts and Media [YCAM]