A Review of Point Set Registration: From Pairwise Registration to Groupwise Registration

Area of Science:

• Computer Vision
• Geometric Modeling
• Computational Geometry

Background:

• Point set registration is crucial for aligning data from different sources.
• Existing methods vary in complexity and accuracy for pairwise and groupwise registration.
• A need exists for efficient and accurate registration algorithms.

Purpose of the Study:

• To provide a comprehensive literature review of point set registration.
• To analyze and compare state-of-the-art modeling methods and algorithms.
• To identify future research directions in point set registration.

Main Methods:

• Literature review of existing point set registration techniques.
• Focus on pairwise and groupwise registration strategies.
• Qualitative and quantitative experimental evaluation of prominent algorithms.

Main Results:

• CPD-GL (Coherent Point Drift - Generalized L-infinity) demonstrated strong performance in pairwise registration.
• JRMPC (Joint Registration and Multi-modal Point Cloud) showed excellent results in groupwise registration.
• Both algorithms outperformed rivals in accuracy and computational complexity on 2D and 3D datasets.

Conclusions:

• CPD-GL and JRMPC represent significant advancements in point set registration.
• These algorithms offer improved accuracy and efficiency for various applications.
• Further research is needed to explore new avenues in point set registration.