Plagiarism meets paraphrasing: insights for the new generation in automatic plagiarism detection

Although paraphrasing is the linguistic mechanism underlying many plagiarism cases, little attention has been paid to its analysis in the framework of automatic plagiarism detection. Therefore, state-of-the-art plagiarism detectors find it difficult to detect cases of paraphrase plagiarism. In this article, we analyse the relationship between paraphrasing and plagiarism, paying special attention to which paraphrase phenomena underlie acts of plagiarism and which of them are detected by plagiarism detection systems. With this aim in mind, we created the P4P corpus, a new resource which uses a paraphrase typology to annotate a subset of the PAN-PC-10 corpus for automatic plagiarism detection. The results of the Second International Competition on Plagiarism Detection were analysed in the light of this annotation. The presented experiments show that (i) more complex paraphrase phenomena and a high density of paraphrase mechanisms make plagiarism detection more difficult, (ii) lexical substitutions are the paraphrase mechanisms used the most when plagiarising, and (iii) paraphrase mechanisms tend to shorten the plagiarized text. For the first time, the paraphrase mechanisms behind plagiarism have been analysed, providing critical insights for the improvement of automatic plagiarism detection systems.

Hand detection in cluttered scene images using Fourier-Mellin invariant features

This paper proposes an automatic hand detection system that combines the Fourier-Mellin Transform along with other computer vision techniques to achieve hand detection in cluttered scene color images. The proposed system uses the Fourier-Mellin Transform as an invariant feature extractor to perform RST invariant hand detection. In a first stage of the system a simple non-adaptive skin color-based image segmentation and an interest point detector based on corners are used in order to identify regions of interest that contains possible matches. A sliding window algorithm is then used to scan the image at different scales performing the FMT calculations only in the previously detected regions of interest and comparing the extracted FM descriptor of the windows with a hand descriptors database obtained from a train image set. The results of the performed experiments suggest the use of Fourier-Mellin invariant features as a promising approach for automatic hand detection.

Hand detection in cluttered scene images using Fourier-Mellin invariant features

This paper proposes an automatic hand detection system that combines the Fourier-Mellin Transform along with other computer vision techniques to achieve hand detection in cluttered scene color images. The proposed system uses the Fourier-Mellin Transform as an invariant feature extractor to perform RST invariant hand detection. In a first stage of the system a simple non-adaptive skin color-based image segmentation and an interest point detector based on corners are used in order to identify regions of interest that contains possible matches. A sliding window algorithm is then used to scan the image at different scales performing the FMT calculations only in the previously detected regions of interest and comparing the extracted FM descriptor of the windows with a hand descriptors database obtained from a train image set. The results of the performed experiments suggest the use of Fourier-Mellin invariant features as a promising approach for automatic hand detection.

The clinical use of quantitative ultrasound (QUS) in the detection and management of osteoporosis

For the detection and management of osteoporosis and osteoporosis-related fractures, quantitative ultrasound (QUS) is emerging as a relatively low-cost and readily accessible alternative to dual-energy X-ray absorptiometry (DXA) measurement of bone mineral density (BMD) in certain circumstances. The following is a brief, but thorough review of the existing literature with respect to the use of QUS in 6 settings: 1) assessing fragility fracture risk; 2) diagnosing osteoporosis; 3) initiating osteoporosis treatment; 4) monitoring osteoporosis treatment; 5) osteoporosis case finding; and 6) quality assurance and control. Many QUS devices exist that are quite different with respect to the parameters they measure and the strength of empirical evidence supporting their use. In general, heel QUS appears to be most tested and most effective. Overall, some, but not all, heel QUS devices are effective assessing fracture risk in some, but not all, populations, the evidence being strongest for Caucasian females over 55 years old. Otherwise, the evidence is fair with respect to certain devices allowing for the accurate diagnosis of likelihood of osteoporosis, and generally fair to poor in terms of QUS use when initiating or monitoring osteoporosis treatment. A reasonable protocol is proposed herein for case-finding purposes, which relies on a combined assessment of clinical risk factors (CR.F) and heel QUS. Finally, several recommendations are made for quality assurance and control.

Spatial filtering for pilot-aided WCDMA systems: a semi-blind subspace approach

This paper proposes a spatial filtering technique forthe reception of pilot-aided multirate multicode direct-sequencecode division multiple access (DS/CDMA) systems such as widebandCDMA (WCDMA). These systems introduce a code-multiplexedpilot sequence that can be used for the estimation of thefilter weights, but the presence of the traffic signal (transmittedat the same time as the pilot sequence) corrupts that estimationand degrades the performance of the filter significantly. This iscaused by the fact that although the traffic and pilot signals areusually designed to be orthogonal, the frequency selectivity of thechannel degrades this orthogonality at hte receiving end. Here,we propose a semi-blind technique that eliminates the self-noisecaused by the code-multiplexing of the pilot. We derive analyticallythe asymptotic performance of both the training-only andthe semi-blind techniques and compare them with the actual simulatedperformance. It is shown, both analytically and via simulation,that high gains can be achieved with respect to training-onlybasedtechniques.

Supervised localimage feature detection

This thesis is about detection of local image features. The research topic belongs to the wider area of object detection, which is a machine vision and pattern recognition problem where an object must be detected (located) in an image. State-of-the-art object detection methods often divide the problem into separate interest point detection and local image description steps, but in this thesis a different technique is used, leading to higher quality image features which enable more precise localization. Instead of using interest point detection the landmark positions are marked manually. Therefore, the quality of the image features is not limited by the interest point detection phase and the learning of image features is simplified. The approach combines both interest point detection and local description into one phase for detection. Computational efficiency of the descriptor is therefore important, leaving out many of the commonly used descriptors as unsuitably heavy. Multiresolution Gabor features has been the main descriptor in this thesis and improving their efficiency is a significant part. Actual image features are formed from descriptors by using a classifierwhich can then recognize similar looking patches in new images. The main classifier is based on Gaussian mixture models. Classifiers are used in one-class classifier configuration where there are only positive training samples without explicit background class. The local image feature detection method has been tested with two freely available face detection databases and a proprietary license plate database. The localization performance was very good in these experiments. Other applications applying the same under-lying techniques are also presented, including object categorization and fault detection.

Development of a systematic computer vision-based method to analyse and compare images of false identity documents for forensic intelligence purposes-Part I: acquisition, calibration and validation issues

Following their detection and seizure by police and border guard authorities, false identity and travel documents are usually scanned, producing digital images. This research investigates the potential of these images to classify false identity documents, highlight links between documents produced by a same modus operandi or same source, and thus support forensic intelligence efforts. Inspired by previous research work about digital images of Ecstasy tablets, a systematic and complete method has been developed to acquire, collect, process and compare images of false identity documents. This first part of the article highlights the critical steps of the method and the development of a prototype that processes regions of interest extracted from images. Acquisition conditions have been fine-tuned in order to optimise reproducibility and comparability of images. Different filters and comparison metrics have been evaluated and the performance of the method has been assessed using two calibration and validation sets of documents, made up of 101 Italian driving licenses and 96 Portuguese passports seized in Switzerland, among which some were known to come from common sources. Results indicate that the use of Hue and Edge filters or their combination to extract profiles from images, and then the comparison of profiles with a Canberra distance-based metric provides the most accurate classification of documents. The method appears also to be quick, efficient and inexpensive. It can be easily operated from remote locations and shared amongst different organisations, which makes it very convenient for future operational applications. The method could serve as a first fast triage method that may help target more resource-intensive profiling methods (based on a visual, physical or chemical examination of documents for instance). Its contribution to forensic intelligence and its application to several sets of false identity documents seized by police and border guards will be developed in a forthcoming article (part II).

Microcalcification evaluation in computer assisted diagnosis for digital mammography

In order to develop applications for z;isual interpretation of medical images, the early detection and evaluation of microcalcifications in digital mammograms is verg important since their presence is oftenassociated with a high incidence of breast cancers. Accurate classification into benign and malignant groups would help improve diagnostic sensitivity as well as reduce the number of unnecessa y biopsies. The challenge here is the selection of the useful features to distinguish benign from malignant micro calcifications. Our purpose in this work is to analyse a microcalcification evaluation method based on a set of shapebased features extracted from the digitised mammography. The segmentation of the microcalcificationsis performed using a fixed-tolerance region growing method to extract boundaries of calcifications with manually selected seed pixels. Taking into account that shapes and sizes of clustered microcalcificationshave been associated with a high risk of carcinoma based on digerent subjective measures, such as whether or not the calcifications are irregular, linear, vermiform, branched, rounded or ring like, our efforts were addressed to obtain a feature set related to the shape. The identification of the pammeters concerning the malignant character of the microcalcifications was performed on a set of 146 mammograms with their real diagnosis known in advance from biopsies. This allowed identifying the following shape-based parameters as the relevant ones: Number of clusters, Number of holes, Area, Feret elongation, Roughness, and Elongation. Further experiments on a set of 70 new mammogmms showed that the performance of the classification scheme is close to the mean performance of three expert radiologists, which allows to consider the proposed method for assisting the diagnosis and encourages to continue the investigation in the senseof adding new features not only related to the shape

Local Features in Image and Video Processing - Object Class Matching and Video Shot Detection

The usage of digital content, such as video clips and images, has increased dramatically during the last decade. Local image features have been applied increasingly in various image and video retrieval applications. This thesis evaluates local features and applies them to image and video processing tasks. The results of the study show that 1) the performance of different local feature detector and descriptor methods vary significantly in object class matching, 2) local features can be applied in image alignment with superior results against the state-of-the-art, 3) the local feature based shot boundary detection method produces promising results, and 4) the local feature based hierarchical video summarization method shows promising new new research direction. In conclusion, this thesis presents the local features as a powerful tool in many applications and the imminent future work should concentrate on improving the quality of the local features.

Generalization of temporal order detection skill learning: two experimental studies of children with dyslexia

The objective of this study was to investigate the phenomenon of learning generalization of a specific skill of auditory temporal processing (temporal order detection) in children with dyslexia. The frequency order discrimination task was applied to children with dyslexia and its effect after training was analyzed in the same trained task and in a different task (duration order discrimination) involving the temporal order discrimination too. During study 1, one group of subjects with dyslexia (N = 12; mean age = 10.9 ± 1.4 years) was trained and compared to a group of untrained dyslexic children (N = 28; mean age = 10.4 ± 2.1 years). In study 2, the performance of a trained dyslexic group (N = 18; mean age = 10.1 ± 2.1 years) was compared at three different times: 2 months before training, at the beginning of training, and at the end of training. Training was carried out for 2 months using a computer program responsible for training frequency ordering skill. In study 1, the trained group showed significant improvement after training only for frequency ordering task compared to the untrained group (P < 0.001). In study 2, the children showed improvement in the last interval in both frequency ordering (P < 0.001) and duration ordering (P = 0.01) tasks. These results showed differences regarding the presence of learning generalization of temporal order detection, since there was generalization of learning in only one of the studies. The presence of methodological differences between the studies, as well as the relationship between trained task and evaluated tasks, are discussed.

Heuristics for the Critical Node Detection Problem in Large Complex Networks

Complex networks have recently attracted a significant amount of research attention due to their ability to model real world phenomena. One important problem often encountered is to limit diffusive processes spread over the network, for example mitigating pandemic disease or computer virus spread. A number of problem formulations have been proposed that aim to solve such problems based on desired network characteristics, such as maintaining the largest network component after node removal. The recently formulated critical node detection problem aims to remove a small subset of vertices from the network such that the residual network has minimum pairwise connectivity. Unfortunately, the problem is NP-hard and also the number of constraints is cubic in number of vertices, making very large scale problems impossible to solve with traditional mathematical programming techniques. Even many approximation algorithm strategies such as dynamic programming, evolutionary algorithms, etc. all are unusable for networks that contain thousands to millions of vertices. A computationally efficient and simple approach is required in such circumstances, but none currently exist. In this thesis, such an algorithm is proposed. The methodology is based on a depth-first search traversal of the network, and a specially designed ranking function that considers information local to each vertex. Due to the variety of network structures, a number of characteristics must be taken into consideration and combined into a single rank that measures the utility of removing each vertex. Since removing a vertex in sequential fashion impacts the network structure, an efficient post-processing algorithm is also proposed to quickly re-rank vertices. Experiments on a range of common complex network models with varying number of vertices are considered, in addition to real world networks. The proposed algorithm, DFSH, is shown to be highly competitive and often outperforms existing strategies such as Google PageRank for minimizing pairwise connectivity.

Impacts and Detection of Design Smells

Les changements sont faits de façon continue dans le code source des logiciels pour prendre en compte les besoins des clients et corriger les fautes. Les changements continus peuvent conduire aux défauts de code et de conception. Les défauts de conception sont des mauvaises solutions à des problèmes récurrents de conception ou d’implémentation, généralement dans le développement orienté objet. Au cours des activités de compréhension et de changement et en raison du temps d’accès au marché, du manque de compréhension, et de leur expérience, les développeurs ne peuvent pas toujours suivre les normes de conception et les techniques de codage comme les patrons de conception. Par conséquent, ils introduisent des défauts de conception dans leurs systèmes. Dans la littérature, plusieurs auteurs ont fait valoir que les défauts de conception rendent les systèmes orientés objet plus difficile à comprendre, plus sujets aux fautes, et plus difficiles à changer que les systèmes sans les défauts de conception. Pourtant, seulement quelques-uns de ces auteurs ont fait une étude empirique sur l’impact des défauts de conception sur la compréhension et aucun d’entre eux n’a étudié l’impact des défauts de conception sur l’effort des développeurs pour corriger les fautes. Dans cette thèse, nous proposons trois principales contributions. La première contribution est une étude empirique pour apporter des preuves de l’impact des défauts de conception sur la compréhension et le changement. Nous concevons et effectuons deux expériences avec 59 sujets, afin d’évaluer l’impact de la composition de deux occurrences de Blob ou deux occurrences de spaghetti code sur la performance des développeurs effectuant des tâches de compréhension et de changement. Nous mesurons la performance des développeurs en utilisant: (1) l’indice de charge de travail de la NASA pour leurs efforts, (2) le temps qu’ils ont passé dans l’accomplissement de leurs tâches, et (3) les pourcentages de bonnes réponses. Les résultats des deux expériences ont montré que deux occurrences de Blob ou de spaghetti code sont un obstacle significatif pour la performance des développeurs lors de tâches de compréhension et de changement. Les résultats obtenus justifient les recherches antérieures sur la spécification et la détection des défauts de conception. Les équipes de développement de logiciels doivent mettre en garde les développeurs contre le nombre élevé d’occurrences de défauts de conception et recommander des refactorisations à chaque étape du processus de développement pour supprimer ces défauts de conception quand c’est possible. Dans la deuxième contribution, nous étudions la relation entre les défauts de conception et les fautes. Nous étudions l’impact de la présence des défauts de conception sur l’effort nécessaire pour corriger les fautes. Nous mesurons l’effort pour corriger les fautes à l’aide de trois indicateurs: (1) la durée de la période de correction, (2) le nombre de champs et méthodes touchés par la correction des fautes et (3) l’entropie des corrections de fautes dans le code-source. Nous menons une étude empirique avec 12 défauts de conception détectés dans 54 versions de quatre systèmes: ArgoUML, Eclipse, Mylyn, et Rhino. Nos résultats ont montré que la durée de la période de correction est plus longue pour les fautes impliquant des classes avec des défauts de conception. En outre, la correction des fautes dans les classes avec des défauts de conception fait changer plus de fichiers, plus les champs et des méthodes. Nous avons également observé que, après la correction d’une faute, le nombre d’occurrences de défauts de conception dans les classes impliquées dans la correction de la faute diminue. Comprendre l’impact des défauts de conception sur l’effort des développeurs pour corriger les fautes est important afin d’aider les équipes de développement pour mieux évaluer et prévoir l’impact de leurs décisions de conception et donc canaliser leurs efforts pour améliorer la qualité de leurs systèmes. Les équipes de développement doivent contrôler et supprimer les défauts de conception de leurs systèmes car ils sont susceptibles d’augmenter les efforts de changement. La troisième contribution concerne la détection des défauts de conception. Pendant les activités de maintenance, il est important de disposer d’un outil capable de détecter les défauts de conception de façon incrémentale et itérative. Ce processus de détection incrémentale et itérative pourrait réduire les coûts, les efforts et les ressources en permettant aux praticiens d’identifier et de prendre en compte les occurrences de défauts de conception comme ils les trouvent lors de la compréhension et des changements. Les chercheurs ont proposé des approches pour détecter les occurrences de défauts de conception, mais ces approches ont actuellement quatre limites: (1) elles nécessitent une connaissance approfondie des défauts de conception, (2) elles ont une précision et un rappel limités, (3) elles ne sont pas itératives et incrémentales et (4) elles ne peuvent pas être appliquées sur des sous-ensembles de systèmes. Pour surmonter ces limitations, nous introduisons SMURF, une nouvelle approche pour détecter les défauts de conception, basé sur une technique d’apprentissage automatique — machines à vecteur de support — et prenant en compte les retours des praticiens. Grâce à une étude empirique portant sur trois systèmes et quatre défauts de conception, nous avons montré que la précision et le rappel de SMURF sont supérieurs à ceux de DETEX et BDTEX lors de la détection des occurrences de défauts de conception. Nous avons également montré que SMURF peut être appliqué à la fois dans les configurations intra-système et inter-système. Enfin, nous avons montré que la précision et le rappel de SMURF sont améliorés quand on prend en compte les retours des praticiens.

Urban Detection From Hyperspectral Images Using Dimension-Reduction Model and Fusion of Multiple Segmentations Based on Stuctural and Textural Features

Ce mémoire de maîtrise présente une nouvelle approche non supervisée pour détecter et segmenter les régions urbaines dans les images hyperspectrales. La méthode proposée n ́ecessite trois étapes. Tout d’abord, afin de réduire le coût calculatoire de notre algorithme, une image couleur du contenu spectral est estimée. A cette fin, une étape de réduction de dimensionalité non-linéaire, basée sur deux critères complémentaires mais contradictoires de bonne visualisation; à savoir la précision et le contraste, est réalisée pour l’affichage couleur de chaque image hyperspectrale. Ensuite, pour discriminer les régions urbaines des régions non urbaines, la seconde étape consiste à extraire quelques caractéristiques discriminantes (et complémentaires) sur cette image hyperspectrale couleur. A cette fin, nous avons extrait une série de paramètres discriminants pour décrire les caractéristiques d’une zone urbaine, principalement composée d’objets manufacturés de formes simples g ́eométriques et régulières. Nous avons utilisé des caractéristiques texturales basées sur les niveaux de gris, la magnitude du gradient ou des paramètres issus de la matrice de co-occurrence combinés avec des caractéristiques structurelles basées sur l’orientation locale du gradient de l’image et la détection locale de segments de droites. Afin de réduire encore la complexité de calcul de notre approche et éviter le problème de la ”malédiction de la dimensionnalité” quand on décide de regrouper des données de dimensions élevées, nous avons décidé de classifier individuellement, dans la dernière étape, chaque caractéristique texturale ou structurelle avec une simple procédure de K-moyennes et ensuite de combiner ces segmentations grossières, obtenues à faible coût, avec un modèle efficace de fusion de cartes de segmentations. Les expérimentations données dans ce rapport montrent que cette stratégie est efficace visuellement et se compare favorablement aux autres méthodes de détection et segmentation de zones urbaines à partir d’images hyperspectrales.