Premium
Beamlets
Author(s) -
Huo Xiaoming
Publication year - 2010
Publication title -
wiley interdisciplinary reviews: computational statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.693
H-Index - 38
eISSN - 1939-0068
pISSN - 1939-5108
DOI - 10.1002/wics.58
Subject(s) - curvilinear coordinates , computer science , statistic , algorithm , data compression , coding (social sciences) , artificial intelligence , noise reduction , pattern recognition (psychology) , data mining , mathematics , statistics , geometry
Beamlets are a data structure for analyzing linear or curvilinear features in two‐dimensional spaces. Beamlets are a collection of dyadically organized line segments, taking a range of positions, orientations, and having various lengths. The key advantage of such a data structure is that it forms a multiscale framework for linear and curvilinear features. This data structure enables new methods in statistical modeling, statistic estimation, denoising, data reduction, coding, and compression. The new methods are both computationally efficient and theoretically optimal. The derived methods are particularly suitable for large‐sized datasets (e.g., images with high resolution). Copyright © 2010 John Wiley & Sons, Inc. This article is categorized under: Applications of Computational Statistics > Signal and Image Processing and Coding