lesterlee89 / klerp Goto Github PK

The file KLERP.h exposes two extremely high performance CPU resize operations, KLERP (bilinear interpolation), and KNERP (nearest neighbor interpolation), both of which use multithreading and various scalar optimizations. KLERP is also written partly in AVX2 so an AVX2-ready CPU is required.

These are state-of-the-art CPU-side interpolators, exceeding OpenCV's implementation in speed by 25-100% while matching its output and capabilities in both interpolation modes.

The magnitude of the speed gain depends on whether you're downscaling or upscaling and by how much.

For even more speed, see my CUDA version, CUDALERP, at https://github.com/komrad36/CUDALERP.

All functionality is contained in the header 'KLERP.h' and has no external dependencies at all.

Note that these are intended for computer vision use (hence the speed) and are designed for grayscale images.

The file 'main.cpp' is an example and speed test driver.