Fast and Compact Graph Cuts for the Boykov-Kolmogorov Algorithm
Researchers have developed a new algorithm, the fast and compact BK (fcBK), that significantly improves the efficiency of computing minimum s-t cuts in graphs. This algorithm achieves a time complexity of O(m|C|), a substantial improvement over previous methods. The fcBK algorithm also utilizes a compact graph representation, enabling it to handle graphs with billions of vertices and edges on standard hardware, and has demonstrated superior performance on benchmark datasets. AI
IMPACT Improves a core computational primitive used in many computer vision and machine learning tasks.