Abstract
Modern rendering libraries provide unprecedented realism, producing real-time photorealistic 3D graphics on commodity hardware. Visual fidelity, however, comes at the cost of increased complexity and difficulty of usage, with many rendering parameters requiring a deep understanding of the pipeline. We propose EasyPBR as an alternative rendering library that strikes a balance between ease-of-use and visual quality. EasyPBR consists of a deferred renderer that implements recent state-of-the-art approaches in physically based rendering. It offers an easy-to-use Python and C++ interface that allows high-quality images to be created in only a few lines of code or directly through a graphical user interface. The user can choose between fully controlling the rendering pipeline or letting EasyPBR automatically infer the best parameters based on the current scene composition. The EasyPBR library can help the community to more easily leverage the power of current GPUs to create realistic images. These can then be used as synthetic data for deep learning or for creating animations for academic purposes.
Introduction
In the field of 3D research, scientists often face a frustrating trade-off between ease of use and visual fidelity. Standard visualization tools like Meshlab or VTK are highly accessible but often produce "flat" or unrealistic results. Conversely, professional-grade engines like Unreal Engine or Blender offer stunning realism but involve steep learning curves, heavy hardware requirements, and cumbersome scripting interfaces that are difficult to integrate into standard machine learning pipelines.
EasyPBR is introduced to solve this "complexity vs. quality" dilemma. It aims to provide "one-click" photorealism by automatically inferring parameters such as lighting and camera clipping, while remaining lightweight enough to be imported as a simple Python module.
Methods
EasyPBR is built on a deferred rendering architecture, which separates geometry processing from shading. This approach allows the system to handle multiple lights and complex post-processing effects efficiently.
PBR Pipeline and G-Buffer
The renderer implements the Cook-Torrance BRDF and Image-Based Lighting (IBL) with split-sum approximations. To optimize performance, EasyPBR utilizes a memory-efficient G-Buffer. Instead of storing world positions for every pixel, it reconstructs them on the fly from the depth map, saving significant bandwidth.
Hybrid Primitive Support
The library supports a variety of geometric primitives:
- Meshes: Standard PBR workflow for high-fidelity surfaces.
- Point Clouds: For datasets lacking surface normals, EasyPBR employs Eye-Dome Lighting (EDL), a non-realistic shading technique that enhances depth perception by darkening edges based on depth discontinuities.
- Surfels: Surface elements (oriented disks) are supported via a weighted splatting technique that accumulates attributes in the G-Buffer before normalization.
Automation and Integration
A core philosophy of EasyPBR is the "Low Barrier" approach. The library includes an "automatic guess" system that sets up a standard 3-point lighting rig and adjusts camera planes based on the scene's bounding box. Furthermore, it provides native interfaces for PyTorch, allowing users to pass tensors directly between neural networks and the visualizer without file conversions. It also shares a common interface with libigl, enabling seamless mesh manipulation and instant rendering.
Experimental Evaluation
The authors evaluated EasyPBR's performance and versatility using several high-resolution assets, including a 23-million-face Goliath statue and 3D scanned head models.
Performance and Efficiency
On an Nvidia RTX 2060, EasyPBR rendered the 23-million-face Goliath mesh in just 6.2ms. It remains competitive with or faster than traditional tools like Meshlab, particularly when handling large geometry. In terms of code efficiency, achieving high-quality shadows and lighting in EasyPBR requires less than 10 lines of Python code, compared to over 150 lines in VTK.
Synthetic Data Generation
The utility of the renderer was demonstrated by training a drone detector on EasyPBR-generated synthetic data. The resulting model successfully generalized to detecting real DJI M100 drones in complex, real-world outdoor environments, proving the library's value for deep learning workflows.
Conclusion and Future Work
While EasyPBR provides a powerful tool for real-time visualization and synthetic data generation, it currently lacks advanced features such as sub-surface scattering for realistic skin and complex transparency handling. Future iterations aim to include cinematic effects like Depth of Field (DoF). While it is not a path-tracer, EasyPBR offers a compelling middle ground for researchers who need high-quality rendering with minimal setup time.
EasyPBR is available on GitHub at: https://github.com/AIS-Bonn/easy_pbr
Summary updated