Adaptive multiscale detection of filamentary structures in a background of uniform random points

被引:30
作者
Arias-Castro, Ery
Donoho, David L.
Huo, Xiaoming
机构
[1] Univ Calif San Diego, Dept Math, La Jolla, CA 92093 USA
[2] Stanford Univ, Dept Stat, Stanford, CA 94305 USA
[3] Georgia Inst Technol, Sch Ind & Syst Engn, Atlanta, GA 30332 USA
关键词
multiscale geometric analysis; pattern recognition; good continuation; Erdos-Renyi laws; runs test; beamlets;
D O I
10.1214/009053605000000787
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
We are given a set of n points that might be uniformly distributed in the unit square [0, 1](2). We wish to test whether the set, although mostly consisting of uniformly scattered points, also contains a small fraction of points sampled from some (a priori unknown) curve with C-alpha-norm bounded by beta. An asymptotic detection threshold exists in this problem; for a constant T_ (alpha, beta) > 0, if the number of points sampled from the curve is smaller than T_(alpha, beta)n(1/(1+alpha)), reliable detection is not possible for large n. We describe a multiscale significant-runs algorithm that can reliably detect concentration of data near a smooth curve, without knowing the smoothness information alpha or 0 in advance, provided that the number of points on the curve exceeds T-*(alpha, beta)n(1/(1+alpha)). This algorithm therefore has an optimal detection threshold, up to a factor T-*/T_. At the heart of our approach is an analysis of the data by counting membership in multiscale multianisotropic strips. The strips will have area 2/n and exhibit a variety of lengths, orientations and anisotropies. The strips are partitioned into anisotropy classes; each class is organized as a directed graph whose vertices all are strips of the same anisotropy and whose edges link such strips to their "good continuations." The point-cloud data are reduced to counts that measure membership in strips. Each anisotropy graph is reduced to a subgraph that consist of strips with significant counts. The algorithm rejects Ho whenever some such subgraph contains a path that connects many consecutive significant counts.
引用
收藏
页码:326 / 349
页数:24
相关论文
共 35 条
[1]  
Abramowicz H, 1997, ADV NEUR IN, V9, P925
[2]  
AHO AV, 1983, DATA STRUCTURES ALGO
[3]   Connect the dots: How many random points can a regular curve pass through? [J].
Arias-Castro, E ;
Donoho, DL ;
Huo, XM ;
Tovey, CA .
ADVANCES IN APPLIED PROBABILITY, 2005, 37 (03) :571-603
[4]   Near-optimal detection of geometric objects by fast multiscale methods [J].
Arias-Castro, E ;
Donoho, DL ;
Huo, XM .
IEEE TRANSACTIONS ON INFORMATION THEORY, 2005, 51 (07) :2402-2425
[5]  
ARIASCASTRO E, 2004, THESIS STANFORD U
[6]   THE ERDOS-RENYI STRONG LAW FOR PATTERN-MATCHING WITH A GIVEN PROPORTION OF MISMATCHES [J].
ARRATIA, R ;
WATERMAN, MS .
ANNALS OF PROBABILITY, 1989, 17 (03) :1152-1169
[7]   Image recognition: Visual grouping, recognition, and learning [J].
Buhmann, JM ;
Malik, J ;
Perona, P .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 1999, 96 (25) :14203-14204
[8]   LOCALIZED RADON TRANSFORM-BASED DETECTION OF SHIP WAKES IN SAR IMAGES [J].
COPELAND, AC ;
RAVICHANDRAN, G ;
TRIVEDI, MM .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 1995, 33 (01) :35-45
[9]   What fMRI has taught us about human vision [J].
Courtney, SM ;
Ungerleider, LG .
CURRENT OPINION IN NEUROBIOLOGY, 1997, 7 (04) :554-561
[10]  
DAVID G, 1993, ANAL UNIFORMLY RECTI