Infinity AI | Shinji Ohmaki
Overview This research proposal explores the potential of generative artificial intelligence (gen-AI) in enhancing and preserving the creative process of contemporary artists, using Shinji Ohmaki’s work as a case study.
Unlike conventional AI models trained on visual patterns and tangible objects, Shinji Ohmaki’s creative practice is rooted in capturing abstract feelings and concepts. He often uses evocative prompts like “death” or “afterlife,” transcending literal representation and pushing the boundaries of how AI understands and interprets human creativity.
Research Questions
Can AI become a collaborative partner in an artist’s creative journey, embracing spontaneity and unpredictability?
How can AI fuel an artist’s creativity, sparking new ideas and concepts throughout the creative process?
Can AI learn to interpret and generate art based on emotional and intentional cues rather than visual references?
How can AI be used to preserve and evolve an artist’s legacy, keeping their work relevant across different time periods and cultural contexts?
Shinji Ohmaki
Shinji Ohmaki is known for creating large-scale installations that inquire into the nature of being and disrupt the somatic senses of viewers. He has presented these ambitious works to high acclaim not only in Japan but also in many other countries across Asia and Europe. In addition, he applies his unparalleled gifts in diverse settings ranging from the performing arts to projects that revitalize local communities, working with a host of collaborators to transform each site.
Engulfed in one of Ohmaki’s spaces, we find ourselves examining our own presence in the world from fresh perspectives. By leaving vestiges of ourselves in these spaces, we arrive at tangible realizations of our own physicality. Whether enveloped in darkness or illuminated by intense light, we are urged to introspectively confront not only our bodies but also our minds and senses.
Evolution of Artistic Legacy
Shinji Ohmaki’s “Infinity Echoes” series, for instance, features motifs inspired by his father’s business, infused with his unique sensibility of color and beauty. This research proposes using AI to reinterpret these motifs across historical and contemporary contexts:
Historical Adaptation: Filtering Shinji’s artwork to fit eras like the Edo period, replacing contemporary patterns with traditional family sigils.
Modernization: Integrating contemporary corporate logos and symbols, adapting Shinji’s aesthetic vision for present-day relevance.
Conceptual Exploration: Envisioning spaces like a department store where each level reflects a distinct time period curated through the evolving language of Shinji’s art.
Digitization of Shinji’s Artworks
The first step in the project involved digitizing and processing of Shinji’s artworks. The digitization process was crucial in capturing the intricate details of his layering technique, which blends color and texture seamlessly.
LoRA Model Training with Kohya_ss
To accurately replicate Shinji Ohmaki’s art style in AI-generated images, we trained a Low-Rank Adaptation (LoRA) model using the Kohya_ss fine-tuning framework. This involved:
Compiling a high-quality dataset from digitized patterns.
Training a LoRA model based on these samples, ensuring output fidelity to Shinji’s unique style.
Evaluating and refining the model through iterative testing and validation.
Stable Diffusion XL Integration via ComfyUI
The trained LoRA model was integrated with Stable Diffusion XL, using VAE Fix to maintain color accuracy and fine detail. This allowed us to generate high-resolution artworks that faithfully adhered to the original aesthetics.
Artisan-Centric UI
A custom UI wrapper was developed on top of ComfyUI making generative AI accessible to artisans without technical expertise. The key features of the UI include:
Prompt-Based Input System – Simplified controls where artisans describe their desired output using natural language.
Craftsmanship-Informed Controls – Custom parameters derived from artistic process, such as artistic voice.
Open-Source Framework - Our technology is built on top of two important open-source frameworks ComfyUI, Kohya_SS and Flow.
Access our growing collection of artworks, crafts, and exclusive products. Be invited to upcoming exhibitions, events, and artist collaborations (20% of your subscription goes directly into funding future artists' exhibitions)
Thank you to the team behind ComfyUI, Kohya_SS and FlowUI for developing open-source tools and AI frameworks.
We would like to extend our heartfelt gratitude to Shinji Ohmaki and his team for supporting our AI training.
For more information, visit Shinji Ohmaki’s Website.
