LIQUID METAL
AI-Driven Character Motion & Abstract Material Study
Case
An internal exploration of abstract material expression, human-like presence, and controlled motion using generative AI image and video workflows.
Date
An exploratory generative AI use case, developed in February 2026.
tools & techniques
Creative AI image generation, abstract colour motif generation, AI-driven motion, and first-frame / last-frame video transitions using Open Source AI models.
Project Overview
LIQUID METAL
Liquid Metal is a self-initiated concept project exploring how generative AI can be used to blend human portraits with evolving, abstract material expressions in motion.
The project combines AI-generated people with fluid, metallic-like colour motifs that interact with the face and body. Rather than treating the abstract elements as static overlays, they are designed to behave as living material, shifting, flowing, and transforming over time.
The focus of the study is not realism, but controlled abstraction, rhythm, and visual continuity across multiple generated clips.
🎥 Watch the video on YouTube: https://youtu.be/j75WIJx8Dx8
🎵 Music created by TOMLIN STUDIO in Logic Pro
Visual Highlights
Project VIDEO & IMAGES
Motion & Transition Exploration
A sequence of still images and short video clips showing variations in abstract material behaviour, colour flow, and character motion.
All base images in Liquid Metal are generated using the Open Source AI model QWEN Image 2512, applied both to the human subject and the abstract colour motifs. These still images are then animated using another Open Source AI video model called WAN 2.2, introducing subtle movement to facial features and abstract material alike.
To create continuity across multiple clips, first-frame / last-frame transitions are used within WAN 2.2. This allows separate generations to be connected into a flowing sequence, where form, colour, and motion evolve naturally from one clip to the next.
The result is a continuous visual experience built from individual AI-generated moments, rather than a single linear animation.
Open Source & Infrastructure
Liquid Metal is generated entirely using Open Source AI models, executed within structured pipelines.
All image and video generation is run on servers based in the EU, ensuring transparency, data control, and alignment with current and upcoming AI regulations.
The project demonstrates how Open Source generative AI can be used for high-quality, expressive audiovisual content, while maintaining control over tooling, infrastructure, and compliance.
IN SUMMERY
Liquid Metal serves as an exploration of:
-
Abstract material behaviour on human subjects
-
Controlled motion in AI-generated portraits
-
Clip-to-clip continuity using first-frame / last-frame logic
-
Open Source AI pipelines for compliant audiovisual production
It positions generative AI not as a one-off image generator, but as a repeatable production system for experimental, character-driven, and material-focused visuals.
products and services
Next-Level Audiovisual Solutions
Generative AI Audiovisuals
Harnessing generative AI to create dynamic and personalized content through advanced work-flow ready generative AI pipelines and through customized generative AI services.
Real-Time Experiences
Utilizing Unreal Engine for interactive, real-world based or 3D-rendered stunning, visuals and immersive interactive experiences that captivate and engage audiences.
Interactive UGC Experiences
Empowering users and players to interact and enjoy while learning. Through playable digital twins or interactive music experiences. In Fortnite or other UGC platforms.
Let's Collaborate on Your Next Vision
We invite you to explore the possibilities of collaboration with us at TOMLIN STUDIO. Whether you’re looking to innovate in digital media using generative AI workflows, create immersive and interactive experiences, or leverage other cutting-edge audiovisual technology, our team is here to bring your vision to life. Reach out today to discuss how we can work together to achieve extraordinary results.