Acoustic features extraction for emotion recognition

In the last decade, the efforts of spoken language processing have achieved significant advances, however, the work with emotional recognition has not progressed so far, and can only achieve 50% to 60% in accuracy. This is because a majority of researchers in this field have focused on the synthesis of emotional speech rather than focusing on automating human emotion recognition. Many research groups have focused on how to improve the performance of the classifier they used for emotion recognition, and few work has been done on data pre-processing, such as the extraction and selection of a set of specifying acoustic features instead of using all the possible ones they had in hand. To work with well-selected acoustic features does not mean to delay the whole job, but this will save much time and resources by removing the irrelative information and reducing the high-dimension data calculation. In this paper, we developed an automatic feature selector based on a RF2TREE algorithm and the traditional C4.5 algorithm. RF2TREE applied here helped us to solve the problems that did not have enough data examples. The ensemble learning technique was applied to enlarge the original data set by building a bagged random forest to generate many virtual examples, and then the new data set was used to train a single decision tree, which selects the most efficient features to represent the speech signals for the emotion recognition. Finally, the output of the selector was a set of specifying acoustic features, produced by RF2TREE and a single decision tree.

DEVELOPMENT OF AN ALGORITHM TO ANALYZE CARTOGRAPHIC FEATURES EXTRACTION METHODS IN DIGITAL IMAGES

The algorithm creates a buffer area around the cartographic features of interest in one of the images and compare it with the other one. During the comparison, the algorithm calculates the number of equals and different points and uses it to calculate the statistical values of the analysis. One calculated statistical value is the correctness, which shows the user the percentage of points that were correctly extracted. Another one is the completeness that shows the percentage of points that really belong to the interest feature. And the third value shows the idea of quality obtained by the extraction method, since that in order to calculate the quality the algorithm uses the correctness and completeness previously calculated. All the performed tests using this algorithm were possible to use the statistical values calculated to represent quantitatively the quality obtained by the extraction method executed. So, it is possible to say that the developed algorithm can be used to analyze extraction methods of cartographic features of interest, since that the results obtained were promising.

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.

Extraction and classification of self-consumable sport video highlights using generic HMM

This paper aims to automatically extract and classify self-consumable sport video highlights. For this purpose, we will emphasize the benefits of using play-break sequences as the effective inputs for HMMbased classifier. HMM is used to model the stochastic pattern of high-level states during specific sport highlights which correspond to the sequence of generic audio-visual measurements extracted from raw video data. This paper uses soccer as the domain study, focusing on the extraction and classification of goal, shot and foul highlights. The experiment work which uses183 play-break sequences from 6 soccer matches will be presented to demonstrate the performance of our proposed scheme.

Automated road pavement marking detection from high resolution aerial images based on Multi-resolution image analysis and anisotropic Gaussian filtering

Road features extraction from remote sensed imagery has been a long-term topic of great interest within the photogrammetry and remote sensing communities for over three decades. The majority of the early work only focused on linear feature detection approaches, with restrictive assumption on image resolution and road appearance. The widely available of high resolution digital aerial images makes it possible to extract sub-road features, e.g. road pavement markings. In this paper, we will focus on the automatic extraction of road lane markings, which are required by various lane-based vehicle applications, such as, autonomous vehicle navigation, and lane departure warning. The proposed approach consists of three phases: i) road centerline extraction from low resolution image, ii) road surface detection in the original image, and iii) pavement marking extraction on the generated road surface. The proposed method was tested on the aerial imagery dataset of the Bruce Highway, Queensland, and the results demonstrate the efficiency of our approach.

Robust speaker verification via fusion of speech and lip modalities

This paper investigates the use of lip information, in conjunction with speech information, for robust speaker verification in the presence of background noise. It has been previously shown in our own work, and in the work of others, that features extracted from a speaker's moving lips hold speaker dependencies which are complementary with speech features. We demonstrate that the fusion of lip and speech information allows for a highly robust speaker verification system which outperforms the performance of either sub-system. We present a new technique for determining the weighting to be applied to each modality so as to optimize the performance of the fused system. Given a correct weighting, lip information is shown to be highly effective for reducing the false acceptance and false rejection error rates in the presence of background noise

Understanding coal seam gas associated water, regulations and strategies for treatment

The coal seam gas (CSG) industry is globally of potentially great importance economically. This study exemplifies the complex relationship between land use and management, groundwater impact and associated water treatment especially in relation to Queensland where a significant increase in the amount of gas extracted over the past 6 years has occurred. In order to effectively manage the environmental impact of the CSG industry it is necessary to appropriately understand the nature of the gas deposits, methods for gas collection, the physicochemical composition of the by-product associated water and the technologies available for water remediation. Australia is mainly considered arid and semi-arid and thus there is a need to not only beneficially reuse water resources but also protect existing ground water reservoirs such as the Great Artesian Basin (GAB). This paper focussed primarily on the Surat Basin located in Queensland and northern New South Wales. The mechanism for CSG formation, relation to local geological features, extraction approach and the potential impact/benefits of associated water was discussed. An outline of the current legislative requirements on physical and chemical properties of associated water in the Surat Basin was also provided, as well as the current treatment technologies used by the major CSG companies. This review was of significance in relation to the formulation of the most appropriate and cost effective management of associated water, while simultaneously preserving existing water resources and the environment.

计算机离线签名鉴定系统研究

介绍了基于VC++开发的离线签名笔迹计算机鉴定系统,能够较好地实现签名笔迹图像多种处理效果,以满足特征提取的不同需要。他可以从复杂签名图像背景下提取出不同颜色签名笔迹,具有方便、快捷、失真小的特点。通过将笔划宽度斜度特征α加入系统进行鉴别,可有效降低识别的错误率,获得了相对较好的鉴别效果。

Palmprint Recognition Based on Subspace Analysis of Gabor Filter Bank

This paper introduces a new technique for palmprint recognition based on Fisher Linear Discriminant Analysis (FLDA) and Gabor filter bank. This method involves convolving a palmprint image with a bank of Gabor filters at different scales and rotations for robust palmprint features extraction. Once these features are extracted, FLDA is applied for dimensionality reduction and class separability. Since the palmprint features are derived from the principal lines, wrinkles and texture along the palm area. One should carefully consider this fact when selecting the appropriate palm region for the feature extraction process in order to enhance recognition accuracy. To address this problem, an improved region of interest (ROI) extraction algorithm is introduced. This algorithm allows for an efficient extraction of the whole palm area by ignoring all the undesirable parts, such as the fingers and background. Experiments have shown that the proposed method yields attractive performances as evidenced by an Equal Error Rate (EER) of 0.03%.

Image to text translation by multi-label classification

This paper presents an image to text translation platform consisting of image segmentation, region features extraction, region blobs clustering, and translation components. A multi-label learning method is suggested for realizing the translation component. Empirical studies show that the predictive performance of the translation component is better than its counterparts when employed a dual-random ensemble multi-label classification algorithm.

Fingertips detection for human interaction system

Fingertips of human hand play an important role in hand-based interaction with computers. Identification of fingertips' positions in hand images is vital for developing a human computer interaction system. This paper proposes a novel method for detecting fingertips of a hand image analyzing the concept of the geometrical structural information of fingers. The research is divided into three parts: First, hand image is segmented for detecting hand, Second, invariant features (curvature zero-crossing points) are extracted from the boundary of the hand, Third, fingertips are detected. Experimental results show that the proposed approach is promising.

Automatic microstructural characterization and classification using artificial intelligence techniques on ultrasound signals

Secondary phases such as Laves and carbides are formed during the final solidification stages of nickel based superalloy coatings deposited during the gas tungsten arc welding cold wire process. However, when aged at high temperatures, other phases can precipitate in the microstructure, like the γ″ and δ phases. This work presents a new application and evaluation of artificial intelligent techniques to classify (the background echo and backscattered) ultrasound signals in order to characterize the microstructure of a Ni-based alloy thermally aged at 650 and 950 °C for 10, 100 and 200 h. The background echo and backscattered ultrasound signals were acquired using transducers with frequencies of 4 and 5 MHz. Thus with the use of features extraction techniques, i.e.; detrended fluctuation analysis and the Hurst method, the accuracy and speed in the classification of the secondary phases from ultrasound signals could be studied. The classifiers under study were the recent optimum-path forest (OPF) and the more traditional support vector machines and Bayesian. The experimental results revealed that the OPF classifier was the fastest and most reliable. In addition, the OPF classifier revealed to be a valid and adequate tool for microstructure characterization through ultrasound signals classification due to its speed, sensitivity, accuracy and reliability. © 2013 Elsevier B.V. All rights reserved.

Improving face recognition with multispectral fusion and support vector machines

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Reconhecimento semi-automático de sinus frontais para identificação humana forense baseado na transformada imagem-floresta e no contexto da forma

Pós-graduação em Ciência da Computação - IBILCE

Desenvolvimento de métodos de processamento e inteligência computacional no ECG ambulatorial

A monitorização ambulatorial do eletrocardiograma (ECG) permite seguir as atividades cotidianas do paciente durante períodos de 24 horas (ou ainda maiores) possibilitando o estudo de casos que pudessem ter episódios arrítmicos fatais. Entretanto, o maior desafio tecnológico que este tipo de monitorização enfrenta é a perda de informação pela presença de ruídos e artefatos quando o paciente se move. A análise do intervalo QT de despolarização e repolarização ventricular do eletrocardiograma superficial é uma técnica não invasiva com um grande valor para a diagnose e prognósticos de cardiopatias e neuropatias, assim como para a predição da morte cardíaca súbita. A análise do desvio padrão do intervalo QT proporciona informação sobre a dispersão (temporal ou espacial) da repolarização ventricular, entretanto a influencia do ruído provoca erros na detecção do final da onda T que são apreciáveis devido ao fato dos valores pequenos do desvio padrão do QT tanto para sujeitos patológicos e quanto para os sãos. O objetivo geral desta tese é melhorar os métodos de processamento do sinal de ECG ambulatorial usando inteligência computacional, especificamente os métodos relacionados com a detecção do final da onda T, e os de reconhecimento morfológico de batimentos que invalidam a análise da variabilidade do intervalo QT. É proposto e validado (em termos de exatidão e precisão) um novo método e algoritmo para estimar o final da onda T baseado no calculo de áreas de trapézios, empregando sinais da base de dados QT da Physionet. O desempenho do método proposto foi testado e comparado com um dos métodos mais usados para detectar o final da onda T: o método baseado no limiar na primeira derivada. O método de inteligência computacional sugerido combina a extração de características usando o método de análise de componentes principais não lineares e a rede neural de tipo perceptron multicamada. O método de áreas de trapézios teve um bom desempenho em condições ruidosas e não depende de nenhum limiar empírico, sendo adequado para situações com níveis de elevados de ruído de banda larga. O método de reconhecimento morfológico de batimentos foi avaliado com sinais ambulatoriais com e sem artefatos pertencentes a bases de dados de prestigio internacional, e mostrou um bom desempenho.