Exploring Fungal Morphology Simulation and Dynamic Light Containment from a Graphics Generation Perspective

Kexin Wang,Ivy He,Jinke Li,Ali Asadipour,Yitong Sun

DOI: https://doi.org/10.48550/arXiv.2409.05171

2024-09-09

Abstract:Fungal simulation and control are considered crucial techniques in Bio-Art creation. However, coding algorithms for reliable fungal simulations have posed significant challenges for artists. This study equates fungal morphology simulation to a two-dimensional graphic time-series generation problem. We propose a zero-coding, neural network-driven cellular automaton. Fungal spread patterns are learned through an image segmentation model and a time-series prediction model, which then supervise the training of neural network cells, enabling them to replicate real-world spreading behaviors. We further implemented dynamic containment of fungal boundaries with lasers. Synchronized with the automaton, the fungus successfully spreads into pre-designed complex shapes in reality.

Graphics,Computer Vision and Pattern Recognition,Human-Computer Interaction,Machine Learning

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to simulate and control fungal morphology and its dynamic light limitation from the perspective of computer - generated graphics. Specifically, the research aims to overcome the programming and technical barriers for artists in traditional fungal simulation algorithms, develop a code - free method to achieve high - fidelity fungal morphology simulation, and be able to guide fungi to grow into pre - designed complex shapes in the real world through lasers.

### Main Problems
1. **Challenges in Fungal Morphology Simulation**:
- Traditional fungal simulation algorithms have relatively high programming and technical barriers for artists.
- It is very difficult to write reliable fungal simulation algorithms, especially when precise control of fungal growth direction and morphology is required.

2. **Controlling Fungal Growth to Achieve Complex Morphologies**:
- How to guide fungi to grow in accordance with pre - designed complex shapes in the real world.
- A method needs to be found so that fungal growth not only conforms to simulation predictions but also enables controllable morphological changes in the actual environment.

### Solutions
To solve the above problems, the paper proposes a simulation pipeline based on a neural - network - driven cellular automaton (NN - cellular automaton) and combines it with dynamic light limitation technology. The specific steps are as follows:

1. **Image Segmentation and Time - Series Prediction**:
- Use the E - ViT (Efficient Vision Transformer) model for image segmentation to capture changes in fungal edges.
- Utilize a customized Temporal Convolutional Network (TCN) to learn fungal diffusion patterns from the segmented images.

2. **Neural Network Training**:
- The TCN model supervises the training of neural network cells, enabling them to autonomously adjust their behavior according to the learned patterns, thereby achieving realistic fungal morphology simulation.

3. **Dynamic Light Limitation**:
- Introduce laser devices to establish dynamic light boundaries in the real world and guide fungi to grow into pre - designed complex shapes.
- The laser devices are synchronized with the simulation environment through an API to ensure that the simulation results are consistent with the actual experiments.

### Key Innovation Points
- **Zero - Coding Implementation**: Users only need to provide a video or image sequence of fungal growth, and the system can automatically generate a high - fidelity fungal morphology simulation without programming knowledge.
- **Efficient Simulation Engine**: Based on the ECS architecture in the Rust language and Vulkan acceleration, efficient cross - platform real - time simulation is achieved.
- **Dynamic Light Control**: Dynamic light limitation of fungal growth is achieved through laser devices, successfully guiding fungi to form complex pre - designed shapes.

### Experimental Verification
The paper verifies the effectiveness of this method through multiple experiments, including random diffusion simulations of different types of fungi, fungal growth control of specific shapes, and practical applications of dynamic light limitation. The experimental results show that this method can not only accurately simulate the natural diffusion patterns of fungi but also effectively control their growth directions and morphologies.

### Conclusion
This research demonstrates the powerful capabilities of combining a neural - network - driven cellular automaton and a dynamic light control strategy, which can accurately simulate and control fungal morphology. This method not only improves the accuracy of fungal simulation but also provides an easy - to - use tool for artists, designers, and researchers, which is suitable for exploration in fields such as bio - art and sustainable materials.