Fatty liver characterization and classification by ultrasound

Steatosis, also known as fatty liver, corresponds to an abnormal retention of lipids within the hepatic cells and reflects an impairment of the normal processes of synthesis and elimination of fat. Several causes may lead to this condition, namely obesity, diabetes, or alcoholism. In this paper an automatic classification algorithm is proposed for the diagnosis of the liver steatosis from ultrasound images. The features are selected in order to catch the same characteristics used by the physicians in the diagnosis of the disease based on visual inspection of the ultrasound images. The algorithm, designed in a Bayesian framework, computes two images: i) a despeckled one, containing the anatomic and echogenic information of the liver, and ii) an image containing only the speckle used to compute the textural features. These images are computed from the estimated RF signal generated by the ultrasound probe where the dynamic range compression performed by the equipment is taken into account. A Bayes classifier, trained with data manually classified by expert clinicians and used as ground truth, reaches an overall accuracy of 95% and a 100% of sensitivity. The main novelties of the method are the estimations of the RF and speckle images which make it possible to accurately compute textural features of the liver parenchyma relevant for the diagnosis.

On the development of an automatic voice pleasantness classification and intensity estimation system

In the last few years, the number of systems and devices that use voice based interaction has grown significantly. For a continued use of these systems, the interface must be reliable and pleasant in order to provide an optimal user experience. However there are currently very few studies that try to evaluate how pleasant is a voice from a perceptual point of view when the final application is a speech based interface. In this paper we present an objective definition for voice pleasantness based on the composition of a representative feature subset and a new automatic voice pleasantness classification and intensity estimation system. Our study is based on a database composed by European Portuguese female voices but the methodology can be extended to male voices or to other languages. In the objective performance evaluation the system achieved a 9.1% error rate for voice pleasantness classification and a 15.7% error rate for voice pleasantness intensity estimation.

Fatty liver automatic diagnosis from ultrasound images

In this paper an automatic classification algorithm is proposed for the diagnosis of the liver steatosis, also known as, fatty liver, from ultrasound images. The features, automatically extracted from the ultrasound images used by the classifier, are basically the ones used by the physicians in the diagnosis of the disease based on visual inspection of the ultrasound images. The main novelty of the method is the utilization of the speckle noise that corrupts the ultrasound images to compute textural features of the liver parenchyma relevant for the diagnosis. The algorithm uses the Bayesian framework to compute a noiseless image, containing anatomic and echogenic information of the liver and a second image containing only the speckle noise used to compute the textural features. The classification results, with the Bayes classifier using manually classified data as ground truth show that the automatic classifier reaches an accuracy of 95% and a 100% of sensitivity.

Automatic musical instrument recognition for multimedia indexing

Trabalho apresentado no âmbito do Mestrado em Engenharia Informática, como requisito parcial para obtenção do grau de Mestre em Engenharia Informática

Automatic Small Bowel Tumor Diagnosis by Using Multi-Scale Wavelet-Based Analysis in Wireless Capsule Endoscopy Images

BACKGROUND: Wireless capsule endoscopy has been introduced as an innovative, non-invasive diagnostic technique for evaluation of the gastrointestinal tract, reaching places where conventional endoscopy is unable to. However, the output of this technique is an 8 hours video, whose analysis by the expert physician is very time consuming. Thus, a computer assisted diagnosis tool to help the physicians to evaluate CE exams faster and more accurately is an important technical challenge and an excellent economical opportunity. METHOD: The set of features proposed in this paper to code textural information is based on statistical modeling of second order textural measures extracted from co-occurrence matrices. To cope with both joint and marginal non-Gaussianity of second order textural measures, higher order moments are used. These statistical moments are taken from the two-dimensional color-scale feature space, where two different scales are considered. Second and higher order moments of textural measures are computed from the co-occurrence matrices computed from images synthesized by the inverse wavelet transform of the wavelet transform containing only the selected scales for the three color channels. The dimensionality of the data is reduced by using Principal Component Analysis. RESULTS: The proposed textural features are then used as the input of a classifier based on artificial neural networks. Classification performances of 93.1% specificity and 93.9% sensitivity are achieved on real data. These promising results open the path towards a deeper study regarding the applicability of this algorithm in computer aided diagnosis systems to assist physicians in their clinical practice.

Spectrum sensing through software defined radio

Dissertação apresentada para obtenção do Grau de Mestre em Engenharia Electrotécnica e de Computadores, pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia

Automatic extraction of concepts from texts and applications

The extraction of relevant terms from texts is an extensively researched task in Text- Mining. Relevant terms have been applied in areas such as Information Retrieval or document clustering and classification. However, relevance has a rather fuzzy nature since the classification of some terms as relevant or not relevant is not consensual. For instance, while words such as "president" and "republic" are generally considered relevant by human evaluators, and words like "the" and "or" are not, terms such as "read" and "finish" gather no consensus about their semantic and informativeness. Concepts, on the other hand, have a less fuzzy nature. Therefore, instead of deciding on the relevance of a term during the extraction phase, as most extractors do, I propose to first extract, from texts, what I have called generic concepts (all concepts) and postpone the decision about relevance for downstream applications, accordingly to their needs. For instance, a keyword extractor may assume that the most relevant keywords are the most frequent concepts on the documents. Moreover, most statistical extractors are incapable of extracting single-word and multi-word expressions using the same methodology. These factors led to the development of the ConceptExtractor, a statistical and language-independent methodology which is explained in Part I of this thesis. In Part II, I will show that the automatic extraction of concepts has great applicability. For instance, for the extraction of keywords from documents, using the Tf-Idf metric only on concepts yields better results than using Tf-Idf without concepts, specially for multi-words. In addition, since concepts can be semantically related to other concepts, this allows us to build implicit document descriptors. These applications led to published work. Finally, I will present some work that, although not published yet, is briefly discussed in this document.

Medium access control design for distributed opportunistic radio networks

Existing wireless networks are characterized by a fixed spectrum assignment policy. However, the scarcity of available spectrum and its inefficient usage demands for a new communication paradigm to exploit the existing spectrum opportunistically. Future Cognitive Radio (CR) devices should be able to sense unoccupied spectrum and will allow the deployment of real opportunistic networks. Still, traditional Physical (PHY) and Medium Access Control (MAC) protocols are not suitable for this new type of networks because they are optimized to operate over fixed assigned frequency bands. Therefore, novel PHY-MAC cross-layer protocols should be developed to cope with the specific features of opportunistic networks. This thesis is mainly focused on the design and evaluation of MAC protocols for Decentralized Cognitive Radio Networks (DCRNs). It starts with a characterization of the spectrum sensing framework based on the Energy-Based Sensing (EBS) technique considering multiple scenarios. Then, guided by the sensing results obtained by the aforementioned technique, we present two novel decentralized CR MAC schemes: the first one designed to operate in single-channel scenarios and the second one to be used in multichannel scenarios. Analytical models for the network goodput, packet service time and individual transmission probability are derived and used to compute the performance of both protocols. Simulation results assess the accuracy of the analytical models as well as the benefits of the proposed CR MAC schemes.

Automatic eye blink detection using consumer web cameras

This research aims to advance blinking detection in the context of work activity. Rather than patients having to attend a clinic, blinking videos can be acquired in a work environment, and further automatically analyzed. Therefore, this paper presents a methodology to perform the automatic detection of eye blink using consumer videos acquired with low-cost web cameras. This methodology includes the detection of the face and eyes of the recorded person, and then it analyzes the low-level features of the eye region to create a quantitative vector. Finally, this vector is classified into one of the two categories considered —open and closed eyes— by using machine learning algorithms. The effectiveness of the proposed methodology was demonstrated since it provides unbiased results with classification errors under 5%

Automatic distinction of Fernando Pessoas’ heteronyms

Text Mining has opened a vast array of possibilities concerning automatic information retrieval from large amounts of text documents. A variety of themes and types of documents can be easily analyzed. More complex features such as those used in Forensic Linguistics can gather deeper understanding from the documents, making possible performing di cult tasks such as author identi cation. In this work we explore the capabilities of simpler Text Mining approaches to author identification of unstructured documents, in particular the ability to distinguish poetic works from two of Fernando Pessoas' heteronyms: Alvaro de Campos and Ricardo Reis. Several processing options were tested and accuracies of 97% were reached, which encourage further developments.

Ontological representation of tumor-node-metastasis classification and an ontology-driven classifier: a study on colorectal cancer

Dissertação de mestrado integrado em Engenharia Biomédica (área de especialização em Informática Médica)

Mise en oeuvre de protocoles spatiaux pour analyser la modulation de processus cognitifs en fonction de la dimension émotionnelle chez le rongeur

Résumé : Emotion et cognition sont deux termes généralement employés pour désigner des processus psychiques de nature opposée. C'est ainsi que les sciences cognitives se sont longtemps efforcées d'écarter la composante «chaude »des processus «froids »qu'elles visaient, si ce n'est pour montrer l'effet dévastateur de la première sur les seconds. Pourtant, les processus cognitifs (de collecte, maintien et utilisation d'information) et émotioAnels (d'activation subjective, physiologique et comportementale face à ce qui est attractif ou aversif) sont indissociables. Par l'approche neuro-éthologique, à l'interface entre le substrat biologique et les manifestations comportementales, nous nous sommes intéressés à une fonction cognitive essentielle, la fonction mnésique, classiquement exprimée chez le rongeur par l'orientation spatiale. Au niveau du substrat, McDonald et White (1993) ont montré la dissociation de trois systèmes de mémoire, avec les rôles de l'hippocampe, du néostriatum et de l'amygdale dans l'encodage des informations respectivement épisodiques, procédurales et émotionnelles. Nous nous sommes penchés sur l'interaction entre ces systèmes en fonction de la dimension émotionnelle par l'éclairage du comportement. L'état émotionnel de l'animal dépend de plusieurs facteurs, que nous avons tenté de contrôler indirectement en comparant leurs effets sur l'acquisition, dans diverses conditions, de la tâche de Morris (qui nécessite la localisation dans un bassin de la position d'une plate-forme submergée), ainsi que sur le style d'exploration de diverses arènes, ouvertes ou fermées, plus ou moins structurées par la présence de tunnels en plexiglas transparent. Nous avons d'abord exploré le rôle d'un composant du système adrénergique dans le rapport à la difficulté et au stress, à l'aide de souris knock-out pour le récepteur à la noradrénaline a-1 B dans un protocole avec 1 ou 4 points de départ dans un bassin partitionné. Ensuite, nous nous sommes penchés, chez le rat, sur les effets de renforcement intermittent dans différentes conditions expérimentales. Dans ces conditions, nous avons également tenté d'analyser en quoi la situation du but dans un paysage donné pouvait interférer avec les effets de certaines formes de stress. Finalement, nous avons interrogé les conséquences de perturbations passées, y compris le renforcement partiel, sur l'organisation des déplacements sur sol sec. Nos résultats montrent la nécessité, pour les souris cont~ô/es dont l'orientation repose sur l'hippocampe, de pouvoir varier les trajectoires, ce qui favoriserait la constitution d'une carte cognitive. Les souris a->B KO s'avèrent plus sensibles au stress et capables de bénéficier de la condition de route qui permet des réponses simples et automatisées, sous-tendues par l'activité du striatum. Chez les rats en bassin 100% renforcé, l'orientation apparaît basée sur l'hippocampe, relayée par le striatum pour le développement d'approches systématiques et rapides, avec réorientation efficace en nouvelle position par réactivation dépendant de l'hippocampe. A 50% de renforcement, on observe un effet du type de déroulement des sessions, transitoirement atténué par la motivation Lorsque les essais s'enchaînent sans pause intrasession, les latences diminuent régulièrement, ce qui suggère une prise en charge possible par des routines S-R dépendant du striatum. L'organisation des mouvements exploratoires apparaît dépendante du niveau d'insécurité, avec différents profils intermédiaires entre la différentiation maximale et la thigmotaxie, qui peuvent être mis en relation avec différents niveaux d'efficacité de l'hippocampe. Ainsi, notre travail encourage à la prise en compte de la dimension émotionnelle comme modulatrice du traitement d'information, tant en phase d'exploration de l'environnement que d'exploitation des connaissances spatiales. Abstract : Emotion and cognition are terms widely used to refer to opposite mental processes. Hence, cognitive science research has for a long time pushed "hot" components away from "cool" targeted processes, except for assessing devastating effects of the former upon the latter. However, cognitive processes (of information collection, preservation, and utilization) and emotional processes (of subjective, physiological, and behavioral activation roue to attraction or aversion) are inseparable. At the crossing between biological substrate and behavioral expression, we studied a chief cognitive function, memory, classically shown in animals through spatial orientation. At the substrate level, McDonald et White (1993) have shown a dissociation between three memory systems, with the hippocampus, neostriatum, and amygdala, encoding respectively episodic, habit, and emotional information. Through the behavior of laboratory rodents, we targeted the interaction between those systems and the emotional axis. The emotional state of an animal depends on different factors, that we tried to check in a roundabout way by the comparison of their effects on acquisition, in a variety of conditions, of the Morris task (in which the location of a hidden platform in a pool is required), as well as on the exploration profile in different apparatus, open-field and closed mazes, more or less organized by clear Plexiglas tunnels. We first tracked the role, under more or less difficult and stressful conditions, of an adrenergic component, with knock-out mice for the a-1 B receptor in a partitioned water maze with 1 or 4 start positions. With rats, we looked for the consequences of partial reinforcement in the water maze in different experimental conditions. In those conditions, we further analyzed how the situation of the goal in the landscape could interfere with the effect of a given stress. At last, we conducted experiments on solid ground, in an open-field and in radial mazes, in order to analyze the organization of spatial behavior following an aversive life event, such as partial reinforcement training in the water maze. Our results emphasize the reliance of normal mice to be able to vary approach trajectories. One of our leading hypotheses is that such strategies are hippocampus-dependent and are best developed for of a "cognitive map like" representation. Alpha-1 B KO mice appear more sensitive to stress and able to take advantage of the route condition allowing simple and automated responses, most likely striatum based. With rats in 100% reinforced water maze, the orientation strategy is predominantly hippocampus dependent (as illustrated by the impairment induced by lesions of this structure) and becomes progressively striatum dependent for the development of systematic and fast successful approaches. Training towards a new platform position requires a hippocampus based strategy. With a 50% reinforcement rate, we found a clear impairment related to intersession disruption, an effect transitorily minimized by motivation enhancement (cold water). When trials are given without intrasession interruption, latencies consistently diminish, suggesting a possibility for striatum dependent stimulus-response routine to occur. The organization of exploratory movements is shown to depend on the level of subjective security, with different intermediary profiles between maximum differentiation and thigmotaxy, which can be considered in parallel with different efficiency levels of the hippocampus dependent strategies. Thus, our work fosters the consideration of emotion as a cognitive treatment modulator, during spatial exploration as well as spatial learning. It leads to a model in which the predominance of hippocampus based exploration is challenged by training conditions of various nature.

Visuo-motor and cognitive procedural learning in children with basal ganglia pathology.

We investigated procedural learning in 18 children with basal ganglia (BG) lesions or dysfunctions of various aetiologies, using a visuo-motor learning test, the Serial Reaction Time (SRT) task, and a cognitive learning test, the Probabilistic Classification Learning (PCL) task. We compared patients with early (<1 year old, n=9), later onset (>6 years old, n=7) or progressive disorder (idiopathic dystonia, n=2). All patients showed deficits in both visuo-motor and cognitive domains, except those with idiopathic dystonia, who displayed preserved classification learning skills. Impairments seem to be independent from the age of onset of pathology. As far as we know, this study is the first to investigate motor and cognitive procedural learning in children with BG damage. Procedural impairments were documented whatever the aetiology of the BG damage/dysfunction and time of pathology onset, thus supporting the claim of very early skill learning development and lack of plasticity in case of damage.

Overview of coded light projection techniques for automatic 3D profiling

Obtaining automatic 3D profile of objects is one of the most important issues in computer vision. With this information, a large number of applications become feasible: from visual inspection of industrial parts to 3D reconstruction of the environment for mobile robots. In order to achieve 3D data, range finders can be used. Coded structured light approach is one of the most widely used techniques to retrieve 3D information of an unknown surface. An overview of the existing techniques as well as a new classification of patterns for structured light sensors is presented. This kind of systems belong to the group of active triangulation method, which are based on projecting a light pattern and imaging the illuminated scene from one or more points of view. Since the patterns are coded, correspondences between points of the image(s) and points of the projected pattern can be easily found. Once correspondences are found, a classical triangulation strategy between camera(s) and projector device leads to the reconstruction of the surface. Advantages and constraints of the different patterns are discussed

Assessment of Severe Apnoea through Voice Analysis, Automatic Speech, and Speaker Recognition Techniques

This study is part of an ongoing collaborative effort between the medical and the signal processing communities to promote research on applying standard Automatic Speech Recognition (ASR) techniques for the automatic diagnosis of patients with severe obstructive sleep apnoea (OSA). Early detection of severe apnoea cases is important so that patients can receive early treatment. Effective ASR-based detection could dramatically cut medical testing time. Working with a carefully designed speech database of healthy and apnoea subjects, we describe an acoustic search for distinctive apnoea voice characteristics. We also study abnormal nasalization in OSA patients by modelling vowels in nasal and nonnasal phonetic contexts using Gaussian Mixture Model (GMM) pattern recognition on speech spectra. Finally, we present experimental findings regarding the discriminative power of GMMs applied to severe apnoea detection. We have achieved an 81% correct classification rate, which is very promising and underpins the interest in this line of inquiry.