NVIDIA HDR Converter GUI

Convert NVIDIA HDR screenshots (JXR format) to JPEG with advanced AI enhancement and tone mapping.

Features

Core Functionality

• Convert NVIDIA JXR (HDR) screenshots to JPEG format
• Advanced HDR tone mapping with multiple algorithms
• AI-powered color enhancement using deep learning models
• Batch processing support
• Live preview with histogram visualization

AI Color Enhancement

• Neural network-based color correction using VGG16, ResNet34, and DenseNet121
• Edge preservation and enhancement using CBAM attention
• Adaptive color balancing
• Intelligent contrast adjustment
• Natural color preservation

Supported Tone Mapping Algorithms

• Adaptive (Default) - Automatically selects the best algorithm based on image content
• Hable
• Reinhard
• Filmic
• ACES
• Uncharted 2

User Interface

• Modern, dark-themed GUI
• Real-time preview with before/after comparison
• Color histogram visualization
• Progress tracking for batch operations
• Customizable enhancement parameters

Requirements

Python Dependencies

torch==2.5.1
torchvision==0.16.1
Pillow==10.2.0
numpy==1.26.4
matplotlib==3.9.3
imagecodecs==2024.9.22
TKinterModernThemes==1.10.4

Installation

1. Clone the repository:

   git clone https://github.com/5ymph0en1x/NVIDIA-HDR-Converter-GUI.git
   cd NVIDIA-HDR-Converter-GUI

2. Install Python dependencies:

   pip install -r requirements.txt

3. Run NHC.py

   python NHC.py

Usage

Single File Conversion

1. Launch the application
2. Select "Single File" mode
3. Choose input JXR file
4. Select output JPEG location
5. Adjust parameters as needed
6. Click "Convert"

Batch Processing

1. Launch the application
2. Select "Folder" mode
3. Choose folder containing JXR files
4. Adjust parameters as needed
5. Click "Convert"

Note: Converted files will be saved in a "Converted_JPGs" subfolder

Parameters

• Tone Map: Select HDR tone mapping algorithm
• Gamma: Adjust gamma correction
• Exposure: Control exposure adjustment
• AI Enhancement: Toggle AI-powered enhancement
• Edge Strength: Adjust edge enhancement intensity

Technical Details

AI Architecture

• Ensemble of VGG16, ResNet34, and DenseNet121 for feature extraction
• Convolutional Block Attention Module (CBAM) for spatial and channel attention
• Edge enhancement block with Sobel filters
• Adaptive color balancing with preservation controls
• Multi-scale feature fusion

Performance Optimization

• Automatic CPU/GPU detection and switching
• Memory-efficient tensor operations
• Multi-threaded batch processing
• Progressive image loading
• Half-precision (FP16) support for GPU processing

Acknowledgments

• Uses PyTorch pretrained models (VGG16, ResNet34, DenseNet121)
• GUI theme based on TKinterModernThemes
• JXR decoding powered by imagecodecs library