waifu2x-ncnn-vulkan – image upscaler and denoiser

waifu2x-ncnn-vulkan is a command-line application for enlarging images and reducing noise using neural network models. It is based on waifu2x and uses the ncnn inference framework to run its models efficiently on compatible graphics hardware.

The program uses Vulkan acceleration, allowing it to work with Intel, AMD and NVIDIA GPUs without requiring the CUDA or Caffe runtime environments. Portable executables are provided with the required models, making the software suitable for use from scripts and other automated workflows.

Although waifu2x was originally developed with anime-style artwork in mind, the application can also be useful for illustrations and other images. The choice of model, noise reduction level and processing settings influences both the quality of the results and the resources required.

This is free and open source software.

Key Features

• Upscale raster images using neural network-based super-resolution models.
• Reduce image noise using configurable denoising levels.
• Use the ncnn framework for neural network inference.
• Accelerate processing through the Vulkan graphics API.
• Support compatible Intel, AMD, NVIDIA and Apple Silicon graphics hardware.
• Provide portable executables with bundled models.
• Run without requiring CUDA or Caffe runtime installations.
• Read and write JPEG, PNG and WebP images.
• Process individual image files or entire directories.
• Select enlargement factors from 1x to 32x.
• Choose noise reduction levels from -1 to 3, including an option to disable denoising.
• Select different waifu2x models through the model directory option.
• Adjust tile size to control GPU memory consumption.
• Use automatic tile-size selection or specify a smaller tile size for memory-constrained systems.
• Select individual GPU devices when more than one is available.
• Configure separate thread counts for image loading, processing and saving.
• Use multiple compatible GPU devices through the relevant processing options.
• Enable test-time augmentation for additional processing at the cost of increased execution time.
• Adjust the output image format independently of the input format.
• Use CPU processing through the appropriate device selection option.

Website: github.com/nihui/waifu2x-ncnn-vulkan
Support:
Developer: nihui
License: MIT License

waifu2x-ncnn-vulkan is written in C++. Learn C++ with our recommended free books and free tutorials.

Related Software

Read our verdict in the software roundup.

LinuxLinks helps you discover the best free and open source software for Linux. Explore our extensive collection of recommended applications, including alternatives to popular proprietary software. We also publish software reviews, Linux tutorials, hardware reviews, free programming books and articles covering the wider open source community. Browse our Linux resources to discover more. Know a useful free and open source Linux application we haven't covered? Suggest a program for inclusion. Has your software been featured on LinuxLinks? Display one of our Featured on LinuxLinks badges on your website or GitHub README.