A free, open-source, automated photo editing app for RAW and JPEG images that applies auto exposure + cinematic LUTs, supporting GPU acceleration, multi-threaded batch processing, and dithering. Think of it as a lightweight, scriptable Lightroom workflow.

As a photographer, I love capturing beautiful moments. As a programmer, I believe repetitive tasks should be automated.

Instead of spending hours in Lightroom manually fixing exposure and applying presets, I wanted a simple open-source tool that:

• 📂 Loads RAW or JPEG images
• 🔆 Auto-corrects exposure & contrast
• 🎨 Applies a cinematic film LUT (.cube)
• 💾 Exports a ready-to-share image

This project is my attempt to combine artistry with automation, making photo editing accessible, fast, and free.

• Auto exposure / contrast stretch
• 3D LUT application (trilinear CPU fallback or GPU-accelerated via PyTorch)
• LUT strength blending (--strength)
• Optional dithering to reduce banding (--no-dither)
• Batch folder processing with multi-threading (--threads)
• Progress bar for large batches
• Supports RAW + JPEG + PNG images
• Automatically organizes output by LUT name

photo-edit-app/
│
├── main.py                 # CLI entry point
├── requirements.txt        # Dependencies
├── examples/
│   ├── sample.jpg
│   ├── cinematic.cube
│   └── cinematic/output.jpg
│
└── photo_edit/
    ├── __init__.py
    ├── pipeline.py         # Image processing pipeline
    ├── lut.py              # LUT load & GPU/CPU apply
    ├── corrections.py      # Auto exposure
    └── io.py               # Image load/save helpers

pip install -r requirements.txt

requirements.txt:

numpy
opencv-python
rawpy
tqdm
scipy
torch
torchvision

python main.py examples/sample.jpg examples/cinematic.cube examples/

• Output will automatically go into a LUT-named folder:

examples/cinematic/sample.jpg

python main.py examples/input_folder examples/cinematic.cube examples/output_folder --threads 8 --strength 0.8

• Processes all supported images in input_folder
• Creates output folder:

examples/output_folder/cinematic/

• Adjustable LUT strength: --strength 0.0..1.0
• Force CPU instead of GPU: --cpu
• Disable dithering: --no-dither

Input → Output (Cinematic LUT)

• JPEG / PNG
• RAW: .nef, .cr2, .arw, .dng, .raf, .rw2

python main.py examples/sample.jpg examples/fuji_fp-100c_alt.cube examples

• Output: examples/fuji_fp-100c_alt/sample.jpg

python main.py examples/input_folder examples/cinematic.cube examples/output_folder --threads 4

• Outputs: examples/output_folder/cinematic/img1.jpg, img2.jpg, …

• GPU acceleration is automatic if CUDA is available.
• Dithering is optional but recommended to reduce color banding on smooth gradients.
• LUTs can be any .cube file (e.g., from GMIC presets or your own).
• Batch processing uses multi-threading to speed up CPU-bound tasks.

MIT License — free to use, modify, and distribute.