How tyhuang98/here-release works
Project type and purpose
• This is a C++ command-line library/demo implementing the “Efficient and Robust Point Cloud Registration via Heuristics-guided Parameter Search” algorithm (T-PAMI 2024).
• Its purpose is to take two or more 3D point-clouds (e.g., PLY files) as input, perform a Branch-and-Bound search with heuristic pruning, and output the rigid transformation aligning them.
File structure
• Top-level files:
• CMakeLists.txt – build orchestration
• demo.cpp – entry point illustrating how to call the library
• / – public headers defining the algorithm’s API and utilities
• registration.h – main registration class/interface
• simu_process.h – simulation or search orchestration
• state_priority_queue.h, state_priority_hashtable.h – the B&B data structures
• utils.h – configuration constants and helper routines
• src/ – implementation of each header
• models/ – sample PLY point-clouds used by the demo
• No web frontend, no database; it’s purely compute-oriented with PCL, Eigen, Boost, OpenMP as external dependencies.
README for additional insights
• Lists external dependencies (Eigen, PCL, Boost, OpenMP).
• Mentions two adjustable parameters: noise_bound and BRANCH_ACCURACY (tunable trade-off between speed and precision).
• Acknowledges that the Branch-and-Bound core is based on an external GORE implementation.