Adaptive feature-specific imaging for recognition of non-Gaussian classes

Pawan K. Baheti; Jun Ke; Mark A. Neifeld

doi:10.1364/AO.48.005225

Adaptive feature-specific imaging for recognition of non-Gaussian classes

Pawan K. Baheti, Jun Ke, Mark A. Neifeld

Electrical and Computer Engineering

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

We present an adaptive feature-specific imaging (AFSI) system for application to an M-class recognition task. The proposed system uses nearest-neighbor-based density estimation to compute the (non-Gaussian) class-conditional densities. We refine the density estimates based on the training data and the knowledge from previous measurements at each step. The projection basis for the AFSI system is also adapted based on the previous measurements at each step. The decision-making process is based on sequential hypothesis testing. We quantify the number of measurements required to achieve a specified probability of error (P_e) and we compare the AFSI system with an adaptive-conventional (ACONV) system. The AFSI system exhibits significant improvement compared to the ACONV system at low signal-to-noise ratio (SNR), and it is shown that, for an M = 4 hypotheses, SNR = -10 dB, and a desired P_e = 10^-2, the AFSI system requires 30 times fewer measurements than the ACONV system. Experimental results validating the AFSI system are presented.

Original language	English (US)
Pages (from-to)	5225-5239
Number of pages	15
Journal	Applied optics
Volume	48
Issue number	28
DOIs	https://doi.org/10.1364/AO.48.005225
State	Published - Oct 1 2009

ASJC Scopus subject areas

Atomic and Molecular Physics, and Optics
Engineering (miscellaneous)
Electrical and Electronic Engineering

Access to Document

10.1364/AO.48.005225

Cite this

@article{dd117ae9684b43b0a79d5d417effce03,

title = "Adaptive feature-specific imaging for recognition of non-Gaussian classes",

abstract = "We present an adaptive feature-specific imaging (AFSI) system for application to an M-class recognition task. The proposed system uses nearest-neighbor-based density estimation to compute the (non-Gaussian) class-conditional densities. We refine the density estimates based on the training data and the knowledge from previous measurements at each step. The projection basis for the AFSI system is also adapted based on the previous measurements at each step. The decision-making process is based on sequential hypothesis testing. We quantify the number of measurements required to achieve a specified probability of error (Pe) and we compare the AFSI system with an adaptive-conventional (ACONV) system. The AFSI system exhibits significant improvement compared to the ACONV system at low signal-to-noise ratio (SNR), and it is shown that, for an M = 4 hypotheses, SNR = -10 dB, and a desired Pe = 10-2, the AFSI system requires 30 times fewer measurements than the ACONV system. Experimental results validating the AFSI system are presented.",

author = "Baheti, {Pawan K.} and Jun Ke and Neifeld, {Mark A.}",

year = "2009",

month = oct,

day = "1",

doi = "10.1364/AO.48.005225",

language = "English (US)",

volume = "48",

pages = "5225--5239",

journal = "Applied optics",

issn = "1559-128X",

publisher = "The Optical Society",

number = "28",

}

TY - JOUR

T1 - Adaptive feature-specific imaging for recognition of non-Gaussian classes

AU - Baheti, Pawan K.

AU - Ke, Jun

AU - Neifeld, Mark A.

PY - 2009/10/1

Y1 - 2009/10/1

N2 - We present an adaptive feature-specific imaging (AFSI) system for application to an M-class recognition task. The proposed system uses nearest-neighbor-based density estimation to compute the (non-Gaussian) class-conditional densities. We refine the density estimates based on the training data and the knowledge from previous measurements at each step. The projection basis for the AFSI system is also adapted based on the previous measurements at each step. The decision-making process is based on sequential hypothesis testing. We quantify the number of measurements required to achieve a specified probability of error (Pe) and we compare the AFSI system with an adaptive-conventional (ACONV) system. The AFSI system exhibits significant improvement compared to the ACONV system at low signal-to-noise ratio (SNR), and it is shown that, for an M = 4 hypotheses, SNR = -10 dB, and a desired Pe = 10-2, the AFSI system requires 30 times fewer measurements than the ACONV system. Experimental results validating the AFSI system are presented.

AB - We present an adaptive feature-specific imaging (AFSI) system for application to an M-class recognition task. The proposed system uses nearest-neighbor-based density estimation to compute the (non-Gaussian) class-conditional densities. We refine the density estimates based on the training data and the knowledge from previous measurements at each step. The projection basis for the AFSI system is also adapted based on the previous measurements at each step. The decision-making process is based on sequential hypothesis testing. We quantify the number of measurements required to achieve a specified probability of error (Pe) and we compare the AFSI system with an adaptive-conventional (ACONV) system. The AFSI system exhibits significant improvement compared to the ACONV system at low signal-to-noise ratio (SNR), and it is shown that, for an M = 4 hypotheses, SNR = -10 dB, and a desired Pe = 10-2, the AFSI system requires 30 times fewer measurements than the ACONV system. Experimental results validating the AFSI system are presented.

UR - http://www.scopus.com/inward/record.url?scp=70349668425&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70349668425&partnerID=8YFLogxK

U2 - 10.1364/AO.48.005225

DO - 10.1364/AO.48.005225

M3 - Article

C2 - 19798360

AN - SCOPUS:70349668425

SN - 1559-128X

VL - 48

SP - 5225

EP - 5239

JO - Applied optics

JF - Applied optics

IS - 28

ER -

Adaptive feature-specific imaging for recognition of non-Gaussian classes

Abstract

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this