作者: J. Derek Tucker , Anuj Srivastava
DOI: 10.1117/12.884537
关键词:
摘要: ABSTRACT In this paper we present a method for clustering and classication of acoustic color data based on statisticalanalysis functions using square-root velocity (SVRF). The convenience the SVRF is that it trans-forms Fisher-Rao metric into standard L 2 metric. As result, formal distance can be calculated usinggeodesic paths. Moreover, allows optimal deformations between to computedfor any two targets allowing robustness measurement error. Using formulation statistical modelscan then constructed principal component analysis model functional variation colordata. Empirical results demonstrate utility improving performance inpattern recognition data.Keywords: color, analysis, pattern recognition, shape sonar imagery 1. INTRODUCTION problem underwater object detection capture has attracted substantialamount attention.