Example Of Pattern Recognition Research Articles

Analysis of human sweat volatiles: An example of pattern recognition in the analysis and interpretation of gas chromatograms

AbstractThe analysis of the volatile components of human body odour is described with a view to establishing how individual identity can be reflected in an odour profile. In order to get good chromatographic replication, a recirculating system was built to load the samples of axillary sweat head‐space into Tenax concentrating traps. The loading temperature and humidity could be controlled and trace volatiles concentrated. The head‐space samples were desorbed from the trap and analysed by GLC. The GC/FID was interfaced with an ATARI ST 1040 computer which stored the output on disc. Pairs of chromatograms were then compared in a pattern‐matching programme which allowed irrelevant differences, such as variations in running time and sample size, to be corrected and then various match parameters to be measured. These were alignment coefficient, profile correlation. Euclidian distance and box car distance. The whole and selected parts of the chromatograms could be matched so that each range of volatiles could be assessed for its contribution to an identity signal.The pattern of sweat volatiles from two pairs of identical twins showed significantly higher match correlations than that of unrelated people. The range of compounds eluting between 80 and 150 C showed the highest match correlations.These techniques could be used to match chromatograms from any source, for example, in drugs and pesticide research, environmental monitoring, food, flavour and perfume analysis or clinical and forensic investigations.

Factory environment and CIM

This contribution on the topic of factory environment and CIM is divided into two parts. The first part deals with some fundamental aspects of intelligent manufacturing systems for the new generation factories. This subject is divided into three blocks: intelligent manufacturing systems, expert systems and intelligent CAE/CAD/CAM. Based on the input-output approach, some basic requirements for IMS definition are given. The second part deals with the problem of the factory environment and CIM, more specifically from historical aspects to the contemporary factory and some conceptual designs aspects of FMS. At the end, an example of structural pattern recognition of FMS is given.

Self induced coherent light resonators with photorefractive KNbO3crystals

Abstract Associative optical memories have attracted increasing interest as they can offer new, all-optical solutions to various problems, for example in pattern recognition. In such memories, stored informations are retrieved from partial or noisy inputs. An optical laser resonator can act as such an associative processor. Using holograms, different images are associated to each of the modes of the optical cavity. Thus a partial input image can force the resonator to oscillate on the closest stored information. Two wave mixing arrangements in photorefractive crystals, which provide very large amplification factors, can be used to overcome the large losses inherent to these resonators.

Effects of binary encoding on pattern recognition and library matching of spectral data

Scott, D.R., 1988. Effects of binary encoding on pattern recognition and library matching of spectral data. Chemometrics and Intelligent Laboratory Systems, 4: 47–63. Binary encoding is frequently employed in pattern recognition studies and matching of unknown against library spectral data. In this study the effect of binary encoding on pattern recognition and library searches is determined using the Hamming and Euclidean distance metrics. The effect on a full intensity spectrum is to compress the total information into a qualitative spectral data vector, the most basic information in the full spectrum. Geometrically, binary encoding of unit normalized spectral data can be visualized as shifting spectral points on the faces of the measurement space hypercube to the corners of a Hamming hypercube. A new classification scheme for comparison of analytical spectra based on their binary encoded spectra is introduced. Quantitative expressions for the effect of binary encoding on general Euclidean distances between spectral points are derived and shown to depend upon their spectral classification. Generally binary encoding increases the interclass distances in pattern recognition and may decrease the intraclass distances. This effect is illustrated with a mass spectral pattern recognition example. The effect of binary encoding on library searches is to produce possible false compound identification in identity searches and to flag spectrally similar compounds in structure searches. A scheme which uses both the Hamming and the Euclidean metrics is proposed for improved library searches. This scheme is illustrated with searches of a small mass spectral library for benzene and p-dioxane spectra.

A model of neural network for spatiotemporal pattern recognition.

A model of neural network to recognize spatiotemporal patterns is presented. The network consists of two kinds of neural cells: P-cells and B-cells. A P-cell generates an impulse responding to more than one impulse and embodies two special functions: short term storage (STS) and heterosynaptic facilitation (HSF). A B-cell generates several impulses with high frequency as soon as it receives an impulse. In recognizing process, an impulse generated by a P-cell represents a recognition of stimulus pattern, and triggers the generation of impulses of a B-cell. Inhibitory impulses with high frequency generated by a B-cell reset the activities of all P-cells in the network. Two examples of spatiotemporal pattern recognition are presented. They are achieved by giving different values to the parameters of the network. In one example, the network recognizes both directional and non-directional patterns. The selectivities to directional and non-directional patterns are realized by only adjusting excitatory synaptic weights of P-cells. In the other example, the network recognizes time series of spatial patterns, where the lengths of the series are not necessarily the same and the transitional speeds of spatial patterns are not always the same. In both examples, the HSF signal controls the total activity of the network, which contributes to exact recognition and error recovery. In the latter example, it plays a role to trigger and execute the recognizing process. Finally, we discuss the correspondence between the model and physiological findings.

On the reducibility of finite Toeplitz matrices - applications in speech analysis and pattern recognition

Using results obtained for centrosymmetric matrices, it is shown that the determinant (characteristic equation) of a Toeplitz matrix can be factored into two (three) products, depending on whether the order N of the matrix is even (odd). This factorization reduces the multiplicative complexity of calculating the determinant by about 75%. Furthermore, by appealing to the specialized class of Toeplitz matrices known as circulants, it is shown that the eigenvalues associated with the ‘even’ and ‘odd’ harmonic eigenvectors interlace regularly in the determinant factors. For general Toeplitz matrices, these eigenvectors arise from the orthogonal polynomials of Grenander and Szegö; whereas, for circulants, the eigenvectors are associated with the Discrete Fourier Transform (DFT). The explicit factorization developed is found to provide an excellent method for assessing the computational complexity of determinant calculation for Toeplitz matrices. For example, it is easily shown that the determinant of a Toeplitz tridiagonal matrix may be computed in exactly 1 2 N additions and 1 2 N multiplications, for N even; also, the determinant of a first order Markovian matrix may be efficiently computed by computing just the determinant of a Markovian matrix of order M, where M is the largest odd divisor of N, for N even. These results in themselves are quite interesting; however, the important point is that the factorization and an orthogonality transformation associated with it seem to return computational attractiveness to the use of a direct approach in the many application areas in which Toeplitz matrices arise. To illustrate this, examples in speech analysis and three-dimensional pattern recognition are discussed in some detail; a rich interpretation now appears to be possible for certain aspects of these problems, due to the factorization results.

Feature selection in human auditory perception

The human ability to organize and make sense out of complex arrays of information is the best example of flexible pattern recognition. Nevertheless, relatively little is known about the processes which underlie these organizational abilities. A model is proposed which describes the organizational rules or criteria employed by human listeners when comparing members of a set of complex sounds. The model assumes that feature selection is based on a Karhunen Loéve expansion of the low-level representations of sound samples. Theoretical and psychophysical analyses were performed on a set of sixteen complex sounds, revealing similar feature representations. It was concluded that the proposed model provides a reasonable first approximation to the organizational rules employed by listeners in a signal comparison task.

Mathematical learning theory: a formalized, axiomatic, abstract approach

An abstract formalism is presented wherein a mathematical learning theory is explored. Numerous examples from the literature are presented. demonstrating how our axiomatic framework formally unifies diverse examples of pattern recognition. Our principal result, the adaptive learning theorem , discusses the expected waiting time of different learning machines. Roughly speaking it states that all algorithms for the induction of boolean relationships between a dependent binary parameter and a finite preassigned set of independent parameters have asymptotically identical average rates of learning. We also give a lower bound for the average learning speed as well as necessary and sufficient conditions for a machine to achieve it. We discuss applications of our methods to classical pattern recognition problems as well as possible application to more complicated research problems. Our principal mathematical innovation is a fruitful correspondence between probability spaces and labeled tree representations.

Pattern recognition of chest X-ray images

AISCR-V2 is a software system for automatic interpretation of chest radiograms. In this paper, an outline of the system is briefly given, and then the method for recognition of rib images is described in detail. The procedure employed here is very unique as a processing method for complex pictures and constitutes the most important part of the system for deriving final decisions correctly. Next, the method for detection of abnormal shadows in the lungs is presented. Finally, experimental results for chest pho-tofluorograms are shown. These results are promising as a first step in automatic diagnosis of chest X-ray films and are very interesting as an example of pattern recognition of extremely complex images.

Laser applications in electronics

The main properties of laser radiation are reviewed and the paper discusses how well practical devices approach the ideal with examples of the latest results which have been achieved. Various applications are then outlined and specific examples in optical communications, holographic stores, spatial filtering and pattern recognition are described.

On finding optimal covers

After a statement of the general problem underlying Quine's methods of simplifying logical expressions, a few examples, and a survey of various approaches to the problem, attention is focussed on the question of how to branch when the straightforward simplication rules give no further progress. An algorithm is suggested that takes advantage of the freedom of choice at the branching step in order to split the given problem into several smaller problems. As the difficulty of the problem grows exponentially with its size, this results in a great saving of effort. Both the hand and computer versions of the algorithm are described since they differ appreciably. The pattern recognition example used to illustrate the paper is chosen as typical of the wide variety of practical questions in which the above general problem arises.