Source Idea
Integrating a physics-based engine with the SceneMaker model will enhance the realism and functional utility of generated 3D scenes, especially in dynamic environments.
View Source Idea →
Files (13)
- README.md
- metadata.json
- requirements.txt
- scripts/run_simulation.py
- src/data_loader.py
- src/evaluate.py
- src/integrate.py
- src/physics_integration.py
- src/train.py
- src/user_study.py
- src/utils.py
- src/vr_interaction.py
- tests/test_integration.py
README Preview
# ScenePhysicsIntegration
## Project Description
This project aims to integrate a physics-based engine with the SceneMaker model to enhance the realism and functional utility of generated 3D scenes, particularly in dynamic environments.
## Research Hypothesis
Integrating a physics-based engine with the SceneMaker model will enhance the realism and functional utility of generated 3D scenes, especially in dynamic environments.
## Implementation Approach
The project will develop an integrated system that combines SceneMaker with a physics engine like Bullet or PhysX. The system will generate scenes involving dynamic interactions and evaluate their realism and accuracy.
## Setup Instructions
1. Clone the repository.
2. Install the required dependencies using `pip install -r requirements.txt`.
3. Set up the environment for VR testing if required.
## Usage Examples
- Run `train.py` to start training and integration.
- Use `evaluate.py` to assess the generated scenes.
## Expected Results
The project aims to produce 3D scenes that not only look realistic but also behave realistically under physical interactions.
## References
- SceneMaker: Open-set 3D Scene Generation with Decoupled De-occlusion and Pose Estimation Model [Paper URL](http://arxiv.org/abs/2512.10957v1)