Phase-Transition Morphogenesis: A Physics-Inspired Framework for Three-Dimensional Generation Through Melting and Resolidification
Abstract
Most contemporary three-dimensional (3D) generative approaches represent shape synthesis as direct geometric prediction, iterative denoising, or optimization of a scene representation. This paper proposes Phase-Transition Morphogenesis (PTM), a conceptual framework for generating 3D geometry through simulated material transformations, particularly melting, liquid redistribution, cooling, and resolidification. Rather than treating the final geometry as the primary object of generation, PTM treats geometry as an outcome of an evolving material state. The proposed framework couples a temperature field, a phase-state field, material transport, solidification dynamics, and a geometry decoder. An initial material configuration undergoes localized phase transitions under controllable thermal and boundary conditions. The resulting material distribution is converted into a 3D surface or another geometric representation. The framework is intended to support the generation of irregular, history-dependent shapes inspired by molten wax, cast metals, and melting or refreezing ice. A differentiable surrogate may be used to learn mappings from conditioning variables to transformation trajectories, while a physics-based solver provides explicit evolution constraints. This paper formalizes the proposed architecture, state variables, governing objectives, training strategy, and evaluation protocol. PTM is presented as a research hypothesis rather than a validated generative model. Its potential contribution is the organization of phase-transition trajectories as a controllable mechanism for 3D shape synthesis. Establishing novelty and effectiveness requires comparison with phase-field simulation, material-point methods, thermomechanical graphics, and existing physics-informed 3D generative approaches.
Related
- Sublimation–Vapor Transport–Redeposition 3D Generation: A Physics-Inspired Framework for Constructing Three-Dimensional Geometry Through Phase-Transition Dynamics
- Freezing-Front-Driven 3D Generation: A Process-Based Framework for Constructing Three-Dimensional Ice Structures Through Phase-Transition Dynamics
- Solidification-Driven 3D Generative Models: A Framework for Temperature-Field-Controlled, Phase-Transition-Based Three-Dimensional Structure Generation
- Heterochronic Uncoupling of Brain and Face Morphogenesis Drives Premature Fusion in Ciliary Transition Zone Mutants
- Splitting-and-Branching 3D Generation: A Recursive Structural Growth Framework for Three-Dimensional Shape Synthesis
BibTeX
@misc{zhang2026phasetransition,
title = {{Phase-Transition Morphogenesis: A Physics-Inspired Framework for Three-Dimensional Generation Through Melting and Resolidification}},
author = {Jincheng Zhang},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23636097},
url = {https://zenodo.org/records/23636097},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Phase-Transition Morphogenesis: A Physics-Inspired Framework for Three-Dimensional Generation Through Melting and Resolidification
AU - Jincheng Zhang
PY - 2026
DA - 2026/10/25
PB - Zenodo
DO - 10.5281/zenodo.23636097
UR - https://zenodo.org/records/23636097
AB - Most contemporary three-dimensional (3D) generative approaches represent shape synthesis as direct geometric prediction, iterative denoising, or optimization of a scene representation. This paper proposes Phase-Transition Morphogenesis (PTM), a conceptual framework for generating 3D geometry through simulated material transformations, particularly melting, liquid redistribution, cooling, and resolidification. Rather than treating the final geometry as the primary object of generation, PTM treats geometry as an outcome of an evolving material state. The proposed framework couples a temperature field, a phase-state field, material transport, solidification dynamics, and a geometry decoder. An initial material configuration undergoes localized phase transitions under controllable thermal and boundary conditions. The resulting material distribution is converted into a 3D surface or another geometric representation. The framework is intended to support the generation of irregular, history-dependent shapes inspired by molten wax, cast metals, and melting or refreezing ice. A differentiable surrogate may be used to learn mappings from conditioning variables to transformation trajectories, while a physics-based solver provides explicit evolution constraints. This paper formalizes the proposed architecture, state variables, governing objectives, training strategy, and evaluation protocol. PTM is presented as a research hypothesis rather than a validated generative model. Its potential contribution is the organization of phase-transition trajectories as a controllable mechanism for 3D shape synthesis. Establishing novelty and effectiveness requires comparison with phase-field simulation, material-point methods, thermomechanical graphics, and existing physics-informed 3D generative approaches.
ER -