OpenCV:hfs
lastmod 2026-11-10created 2023-10-252 edits
HFS (Hierarchical Feature Selection) is an algorithm that segments an image by grouping similar regions together, implemented as hfs among OpenCV's contrib modules. It first divides the image finely into SLIC superpixels, then hierarchically selects features such as color and texture to merge adjacent regions step by step, achieving near-real-time speed. A GPU (CUDA) accelerated implementation is also provided.
HFS: Hierarchical Feature Selection for Efficient Image Segmentation