This research addresses the problem of temporal pattern discovery in real-valued, multivariate sensor data. Several algorithms were developed, and subsequent evaluation demonstrates that they can efficiently and accurately discover unknown recurring patterns in time series data taken from many different domains. Different data representations and motif models were investigated in order to design an algorithm with an improved balance between run-time and detection accuracy. The different data representations are used to quickly filter large data sets in order to detect potential patterns that form the basis of a more detailed analysis. The representations include global discretization, which can be efficiently analyzed using a suffix tree, local discretization with a corresponding random projection algorithm for locating similar pairs of subsequences, and a density-based detection method that operates on the original, real-valued data. In addition, a new variation of the multivariate motif discovery problem is proposed in which each pattern may span only a subset of the input features. An algorithm that can efficiently discover such "subdimensional" patterns was developed and evaluated. The discovery algorithms are evaluated by measuring the detection accuracy of discovered patterns relative to a set of expected patterns for each data set. The data sets used for evaluation are drawn from a variety of domains including speech, on-body inertial sensors, music, American Sign Language video, and GPS tracks.

The problem of sensing a medium by several sensors and retrieving interesting features is a very general one. The basic framework of the problem is generally the same for applications from MRI, tomography, Radar SAR imaging to subsurface imaging, even though the data acquisition processes, sensing geometries and sensed properties are different. In this thesis we introduced a new perspective to the problem of remote sensing and information retrieval by studying the problem of subsurface imaging using GPR and seismic sensors. We have shown that if the sensed medium is sparse in some domain then it can be imaged using many fewer measurements than required by the standard methods. This leads to much lower data acquisition times and better images representing the medium. We have used the ideas from Compressive Sensing, which show that a small number of random measurements about a signal is sufficient to completely characterize it, if the signal is sparse or compressible in some domain. Although we have applied our ideas to the subsurface imaging problem, our results are general and can be extended to other remote sensing applications. A second objective in remote sensing is information retrieval which involves searching for important features in the computed image of the medium. In this thesis we focus on detecting buried structures like pipes, and tunnels in computed GPR or seismic images. The problem of finding these structures in high clutter and noise conditions, and finding them faster than the standard shape detecting methods like the Hough transform is analyzed. One of the most important contributions of this thesis is, where the sensing and the information retrieval stages are unified in a single framework using compressive sensing. Instead of taking lots of standard measurements to compute the image of the medium and search the necessary information in the computed image, a much smaller number of measurements as random projections are taken. The data acquisition and information retrieval stages are unified by using a data model dictionary that connects the information to the sensor data.

[Truncated abstract] The field of pattern recognition aims to develop algorithms and computer programs that can learn patterns from data, where learning encompasses the problems of recognition, representation, classification and prediction. Syntactic pattern recognition recognises that patterns may be hierarchically structured. Formal language theory is an example of a syntactic approach, and is used extensively in computer languages and speech processing. However, the underlying structure of language and speech is strictly one-dimensional. The application of syntactic pattern recognition to the analysis of images requires an extension of formal language theory. Thus, this thesis extends and generalises formal language theory to apply to data that have possibly multi-dimensional underlying structure and also hierarchic structure . . . As in the case for curves, shapes are modelled as a sequence of local relationships between the curves, and these are estimated using a training sample. Syntactic square detection was extremely successful – detecting 100% of squares in images containing only a single square, and over 50% of the squares in images containing ten squares highly likely to be partially or severely occluded. The detection and classification of polygons was successful, despite a tendency for occluded squares and rectangles to be confused. The algorithm also peformed well on real images containing fish. The success of the syntactic approaches for detecting edges, detecting curves and detecting, classifying and counting occluded shapes is evidence of the potential of syntactic models.

Thesis (Ph.D.)--University of Adelaide, Dept. of Electrical and Electronic Engineering, 1997. / Bibliography: leaves 146-159.

Thesis (M.Comp.Sc.)--University of Wollongong, 2008. / Typescript. Includes bibliographical references: leaf 88-98.

Dissertation (Ph.D.)--The University of North Carolina at Greensboro, 2008. / Title from PDF t.p. (viewed Aug. 10, 2009). Directed by Roy Stine; submitted to the Dept. of Geography. Includes bibliographical references (p. 154-165).

Thesis (Ph.D.) - Carleton University, 2007. / Includes bibliographical references (p. 219-233). Also available in electronic format on the Internet.

Thesis (MEng)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: We present the development of a pattern recognition system which is capable of classifying different Spoken Variants (SVs) of South African English (SAE) using a transcriptionless speech recognition approach. Spoken Variants (SVs) allow us to unify the linguistic concepts of accent and dialect from a pattern recognition viewpoint. The need for the SAE SV classification system arose from the multi-linguality requirement for South African speech recognition applications and the costs involved in developing such applications. / AFRIKAANSE OPSOMMING: Ons beskryf die ontwikkeling van 'n patroon herkenning stelsel wat in staat is om verskillende Gesproke Variante (GVe) van Suid Afrikaanse Engels (SAE) te klassifiseer met behulp van 'n transkripsielose spraak herkenning metode. Gesproke Variante (GVe) stel ons in staat om die taalkundige begrippe van aksent en dialek te verenig vanuit 'n patroon her kenning oogpunt. Die behoefte aan 'n SAE GV klassifikasie stelsel het ontstaan uit die meertaligheid vereiste vir Suid Afrikaanse spraak herkenning stelsels en die koste verbonde aan die ontwikkeling van sodanige stelsels.

Cervical cancer is one of the main causes of death by disease worldwide. In Peru, it holds the first place in frequency and represents 8% of deaths caused by sickness. To detect the disease in the early stages, one of the most used screening tests is the cervix Papanicolaou test. Currently, digital images are increasingly being used to improve Pap test efficiency. This work develops an algorithm based on adaptive thresholds, which will be used in Pap smear assisted quality control software. The first stage of the method is a pre-processing step, in which noise and background removal is done. Next, a block is segmented for each one of the points selected as not background, and a local threshold per block is calculated to search for cell nuclei. If a nucleus is detected, an artifact rejection follows, where only cell nuclei and inflammatory cells are left for the doctors to interpret. The method was validated with a set of 55 images containing 2317 cells. The algorithm successfully recognized 92.3% of the total nuclei in all images collected. / Revisón por pares

Le système d’identification d’iris est considéré comme l’une des meilleures technologies biométriques. Toutefois, des problèmes liés à la segmentation de l’iris et à la normalisation de la texture de l’iris sont généralement signalés comme principales origines des reconnaissances incorrectes. Dans notre travail, trois contributions principales sont proposées pour améliorer le système d’identification d’iris. Une nouvelle méthode de segmentation est développée. Elle détecte la frontière externe de l’iris par un contour circulaire et la pupille, d’une manière précise, à l’aide d’un modèle de contour actif. Ensuite, une nouvelle méthode de normalisation est proposée. Elle assure une représentation plus robuste et un meilleur échantillonnage de la texture de l’iris comparée aux méthodes traditionnelles. Enfin en utilisant le système d’identification d’iris proposé, la localisation des caractéristiques discriminantes dans une région d’iris est identifiée. Nous vérifions que l’information la plus importante de la région de l’iris se trouve à proximité de la pupille et que la capacité discriminante de la texture diminue avec la distance à la pupille. Les méthodes de segmentation et de normalisation développées sont testées et comparées à un système de référence sur une base de données contenant 2639 images d’iris. Une amélioration des performances de reconnaissance valide l’efficacité du système proposé / Iris identification system is considered among the best biometric technologies. However problems related to the segmentation of the iris and to the normalization of iris templates are generally reported and induce loss of recognition performance. In this work three main contributions are made to the progress of the iris identification system. A new segmentation method is developed. It approximates the outer iris boundary with a circle and segments accurately the inner boundary of the iris by use of an active contour model. Next, a new normalization method is proposed. It leads to a more robust characterization and a better sampling of iris textures compared to traditional normalization methods. Finally using the proposed iris identification system, the location of discriminant characteristics along iris templates is identified. It appears that the most discriminant iris characteristics are located in inner regions of the iris (close to the pupil boundary) and that the discriminant capabilities of these characteristics decreases as outer regions of the iris are considered. The developed segmentation and normalization methods are tested and compared to a reference iris identification system over a database of 2639 iris images. Improvement in recognition performance validates the effectiveness of the proposed system