Spectrum analysis A modern perspective

This Article covers the main topic of Spectrum analysis and will use as a base to my main research.

Abstract:

A summary of many of the new techniques developed in the last two decades for spectrum analysis of discrete time series is presented in this tutorial. An examination of the underlying time series model assumed by each technique serves as the common basis for understanding the differences among the various spectrum analysis approaches. Techniques discussed include the classical periodogram, classical Blackman-Tukey, autoregressive (maximum entropy), moving average, autotegressive-moving average, maximum likelihood, Prony, and Pisarenko methods. A summary table in the text provides a concise overview for all methods, including key references and appropriate equations for computation of each spectral estimate.

Link to Article:

Author(s) Kay, S.M. University of Rhode Island, Kingston, RI Marple, S.L., Jr.

ISSN : 0018-9219Date of Current Version :28 June 2005Digital Object Identifier : 10.1109/PROC.1981.12184

Measurement of sugar content of white vinegars using VIS/near-infrared.

This has all the attributes of my final keywords , and abstract contains my overall goals matching highly with my research. But not cited that much.

Abstract:

Visible and near infrared (VIS/NIR) spectroscopy combined with different calibration models was applied to predict the sugar content of white vinegars. The calibration set was composed of 240 samples, whereas 80 samples in the validation set. Partial least squares (PLS) models with or without pretreatments were developed and certain latent variables (LVs) were extracted by PLS analysis. The selected LVs were used as the inputs of BP neural network (BPNN) model. Finally, three models were developed. The prediction results indicated that PLS model with no pretreatment was better than that with pretreatments, and the best performance was obtained by BPNN model. The correlation coefficient, RMSEP and bias for validation set by BPNN model were 0.995, 0.135 and 0.035, respectively. The overall results indicated that VIS/NIR spectroscopy could be used as an alternative approach for the prediction of sugar content, and the BPNN models achieved the optimal prediction accuracy.

Link to Article:

Page(s):1311 - 1316

Conference Location :Kunming

Print ISBN: 978-1-4244-2095-7

Complete discrete 2-D Gabor transforms by neural networks for image analysis and compression

This is what I may use to convert my spectral image , well at least a foundation for it


Abstract:

A three-layered neural network is described for transforming two-dimensional discrete signals into generalized nonorthogonal 2-D Gabor representations for image analysis, segmentation, and compression. These transforms are conjoint spatial/spectral representations, which provide a complete image description in terms of locally windowed 2-D spectral coordinates embedded within global 2-D spatial coordinates. In the present neural network approach, based on interlaminar interactions involving two layers with fixed weights and one layer with adjustable weights, the network finds coefficients for complete conjoint 2-D Gabor transforms without restrictive conditions. In wavelet expansions based on a biologically inspired log-polar ensemble of dilations, rotations, and translations of a single underlying 2-D Gabor wavelet template, image compression is illustrated with ratios up to 20:1. Also demonstrated is image segmentation based on the clustering of coefficients in the complete 2-D Gabor transform

Link to Article:

Published in:
Acoustics, Speech and Signal Processing, IEEE Transactions on  (Volume:36 ,  Issue: 7 )

ISSN :0096-3518

Date of Current Version :06 August 2002Digital Object Identifier :10.1109/29.1644

Colorimeter and spectrcopy Digital Sensing using ANNS

Uses BP ANNS to analysis Tea , Using Near Infrared So a complete fit for my research fits my check-list very good except for citations.

Discrimination of varieties of tea using near infrared spectroscopy by principal component analysis and BP model Yong He, Xiaoli Li, Xunfei Deng College of Biosystems Engineering and Food Science, Zhejiang University, 310029, Hangzhou, China.

http://www.aseanfood.info/Articles/11017701.pdf

 
Digital Object Identifier: 10.1109/ICCIAS.2006.295409 

A review of Bayesian neural networks with an application to near infrared spectroscopy

May be of use but very heavy maths will have to review better but matches my check-list well , has my exact keywords but will look at different Neural Nets and see what one will be the best for me.

link to Journal:

Abstract:

MacKay's (1992) Bayesian framework for backpropagation is a practical and powerful means to improve the generalization ability of neural networks. It is based on a Gaussian approximation to the posterior weight distribution. The framework is extended, reviewed, and demonstrated in a pedagogical way. The notation is simplified using the ordinary weight decay parameter, and a detailed and explicit procedure for adjusting several weight decay parameters is given. Bayesian backprop is applied in the prediction of fat content in minced meat from near infrared spectra. It outperforms “early stopping” as well as quadratic regression. The evidence of a committee of differently trained networks is computed, and the corresponding improved generalization is verified. The error bars on the predictions of the fat content are computed. There are three contributors: The random noise, the uncertainty in the weights, and the deviation among the committee members. The Bayesian framework is compared to Moody's GPE (1992). Finally, MacKay and Neal's automatic relevance determination, in which the weight decay parameters depend on the input number, is applied to the data with improved results

Published in:
Neural Networks, IEEE Transactions on  (Volume:7 ,  Issue: 1 )

Date of Publication: Jan 1996

Page(s):

56 - 72

ISSN :

1045-9227

INSPEC Accession Number:

5181044

Digital Object Identifier :

10.1109/72.478392

Date of Current Version :

06 August 2002

Issue Date :

Jan 1996

Sponsored by :

IEEE Computational Intelligence Society

Statistical pattern recognition: a review

May give me a different perspective to my Neural net:


Abstract:

The primary goal of pattern recognition is supervised or unsupervised classification. Among the various frameworks in which pattern recognition has been traditionally formulated, the statistical approach has been most intensively studied and used in practice. More recently, neural network techniques and methods imported from statistical learning theory have been receiving increasing attention. The design of a recognition system requires careful attention to the following issues: definition of pattern classes, sensing environment, pattern representation, feature extraction and selection, cluster analysis, classifier design and learning, selection of training and test samples, and performance evaluation. In spite of almost 50 years of research and development in this field, the general problem of recognizing complex patterns with arbitrary orientation, location, and scale remains unsolved. New and emerging applications, such as data mining, web searching, retrieval of multimedia data, face recognition, and cursive handwriting recognition, require robust and efficient pattern recognition techniques. The objective of this review paper is to summarize and compare some of the well-known methods used in various stages of a pattern recognition system and identify research topics and applications which are at the forefront of this exciting and challenging field

Link to:

Published in: Pattern Analysis and Machine Intelligence, IEEE Transactions on  (Volume:22 ,  Issue: 1 )
Date of Publication: Jan 2000Page(s):4 - 37

ISSN :

0162-8828

INSPEC Accession Number:

6525225

Digital Object Identifier :

10.1109/34.824819

Date of Current Version :

06 August 2002

Issue Date :

Jan 2000

Sponsored by :

IEEE Computer Society