🎨 PyBlender Rendering Pipeline

A robust, GPU-accelerated Python pipeline for rendering 3D point cloud and mesh comparisons using Blender's Cycles renderer.

---

📖 Overview

PyBlender is designed to visualize 3D point cloud and mesh denoising/reconstruction results. It computes distance-based errors between ground truth and reconstructed point clouds, applies colormaps (e.g., viridis, turbo) to represent these distances, and renders high-quality images using Blender Cycles.

Key features include:

• Fast GPU Rendering: OptiX denoising on the GPU for ~8x faster renders.
• Automated Workflows: Ready-to-run Jupyter Notebooks for Google Colab and Kaggle.
• Cloud Integration: Async uploads directly to Google Cloud Storage (GCS) and Weights & Biases (WandB) tracking.
• Extensive Colormap Support: Compare point clouds using 168+ matplotlib colormaps.

---

🏗️ System Architecture

flowchart LR
    A["Raw Data\n(.xyz / .ply)"] --> B{"Execution Environment"}
    
    subgraph Environments [Rendering Engines]
        B -->|Google Colab| C["Colab_Script/\nSingle/Multi-Colormap"]
        B -->|Kaggle (TPU/CPU)| D["Kaggel_Script/\nHigh Performance"]
        B -->|Local PC| E["Data.py + Local Scripts"]
    end
    
    C --> F["Blender Cycles\n(Python 3.10)"]
    D --> F
    E --> F
    
    F -->|GPU Rendering| G["Outputs\n(.png / .blend)"]
    G --> H["Google Drive / GCS"]
    G --> I["Colormap Catalog\nWebsite"]

Loading

---

📂 Repository Structure

PyBlender/
├── 📁 BlenderToolbox/              # Cloned dependency for Blender Python utilities
├── 📁 Colab_Script/                # Ready-to-use Google Colab notebooks
├── 📁 Kaggel_Script/               # High-performance Kaggle notebooks with GCS/WandB
├── 📁 Colormap_Catloge_Website/    # Vite/React app for browsing rendered colormaps
├── 📁 Data/                        # 📦 Point clouds & meshes (Ignored in git, download required)
├── 📁 Deprecate/                   # Legacy Python scripts (Colored_PC, etc.)
├── 📁 Tools/                       # Utility scripts (Clean notebooks, extract PNGs)
├── 📄 Data.py                      # Script to download asset data from Google Drive
├── 📄 requirements.txt             # Python dependencies
└── 📄 README.md                    # You are here!

For more details, check the README.md inside each respective directory.

---

🚀 Quick Start Guide

1. Requirements

• Blender 4.0+ installed on your system.
• Python 3.10+ (Recommended: use a virtual environment).

2. Installation

Clone the repository and set up your virtual environment:

git clone https://github.com/Mahendra2409/PyBlender.git
cd PyBlender

# Create and activate virtual environment (Windows)
python -m venv .venv
.venv\Scripts\activate

# (Linux / Mac)
# source .venv/bin/activate

Install the required Python dependencies:

pip install -r requirements.txt

Clone the required BlenderToolbox repository:

git clone https://github.com/HTDerekLiu/BlenderToolbox.git

3. Download the Data

Because 3D point cloud data is large, it is not stored in this repository. You need to download it directly. Run the download script:

python Data.py

(Alternatively, you can manually download the data from the link inside Data/README.md)

4. Running Scripts

You are now ready to run scripts!

• For Local runs, navigate to the specific script directories or explore the Tools/.
• For Cloud rendering, explore the Colab Scripts or Kaggle Scripts directories and run the provided notebooks.

---

Note

Are you looking for the legacy local rendering scripts? They have been moved to the Deprecate/ directory in favor of cloud-optimized Colab and Kaggle pipelines.