Recognition of Facial Expressions with Autoencoders and Convolutional-Nets

Les humains communiquent via différents types de canaux: les mots, la voix, les gestes du corps, des émotions, etc. Pour cette raison, un ordinateur doit percevoir ces divers canaux de communication pour pouvoir interagir intelligemment avec les humains, par exemple en faisant usage de microphones et de webcams. Dans cette thèse, nous nous intéressons à déterminer les émotions humaines à partir d’images ou de vidéo de visages afin d’ensuite utiliser ces informations dans différents domaines d’applications. Ce mémoire débute par une brève introduction à l'apprentissage machine en s’attardant aux modèles et algorithmes que nous avons utilisés tels que les perceptrons multicouches, réseaux de neurones à convolution et autoencodeurs. Elle présente ensuite les résultats de l'application de ces modèles sur plusieurs ensembles de données d'expressions et émotions faciales. Nous nous concentrons sur l'étude des différents types d’autoencodeurs (autoencodeur débruitant, autoencodeur contractant, etc) afin de révéler certaines de leurs limitations, comme la possibilité d'obtenir de la coadaptation entre les filtres ou encore d’obtenir une courbe spectrale trop lisse, et étudions de nouvelles idées pour répondre à ces problèmes. Nous proposons également une nouvelle approche pour surmonter une limite des autoencodeurs traditionnellement entrainés de façon purement non-supervisée, c'est-à-dire sans utiliser aucune connaissance de la tâche que nous voulons finalement résoudre (comme la prévision des étiquettes de classe) en développant un nouveau critère d'apprentissage semi-supervisé qui exploite un faible nombre de données étiquetées en combinaison avec une grande quantité de données non-étiquetées afin d'apprendre une représentation adaptée à la tâche de classification, et d'obtenir une meilleure performance de classification. Finalement, nous décrivons le fonctionnement général de notre système de détection d'émotions et proposons de nouvelles idées pouvant mener à de futurs travaux.

Entre les bonnes intentions et les effets pervers : l’utilisation des sports de combat comme levier d’intervention auprès des jeunes à risque

L’objectif de cette étude est de décrire et comprendre l’impact d’un programme de sport-étude mis sur pied pour les jeunes à risque sur la vie des participants. Le programme à l’étude a été mis sur pied à Montréal afin de maximiser l’intégration sociale de jeunes à risque de s’engager, ou étant déjà engagés dans un mode de vie déviant. Le programme mise sur les sports de combat comme levier d’intervention auprès de ces jeunes. Les participants sont invités à suivre une formation scolaire aux adultes en matinée et des entraînements en sports de combat en après-midi. L’objectif de ce programme est de les outiller pour qu’ils intègrent la société efficacement. Une méthodologie qualitative a été choisie afin de mener cette étude à terme. Une période d’observation participante et des entretiens semi-dirigés avec différents acteurs du programme ont permis de bien saisir la dynamique à l’intérieur de celui-ci et de déterminer quels sont les impacts de la participation au programme sur la vie des jeunes. Le programme génère des effets mixtes sur la vie des participants: 1) quelques uns y vivent des changements positifs, notamment par rapport à l’estime de soi et l’autodiscipline, 2) d’autres ne tirent pas de bénéfices particuliers de leur passage dans le programme et 3) certains développent un sentiment d’indifférence, d’apathie à la suite de leur passage dans le programme, notamment induit par la dynamique présente au sein de celui-ci. Il appert que le modèle logique du programme n’a pas été fidèlement respecté lors de sa mise en place. Un mauvais appariement entre la clientèle et le programme ainsi qu’une implantation déficiente du modèle logique sont à la base des effets pervers induits par celui-ci.

Development and Evaluation of Blind Identification Techniques for Nonlinear Systems

Identification and Control of Non‐linear dynamical systems are challenging problems to the control engineers.The topic is equally relevant in communication,weather prediction ,bio medical systems and even in social systems,where nonlinearity is an integral part of the system behavior.Most of the real world systems are nonlinear in nature and wide applications are there for nonlinear system identification/modeling.The basic approach in analyzing the nonlinear systems is to build a model from known behavior manifest in the form of system output.The problem of modeling boils down to computing a suitably parameterized model,representing the process.The parameters of the model are adjusted to optimize a performanace function,based on error between the given process output and identified process/model output.While the linear system identification is well established with many classical approaches,most of those methods cannot be directly applied for nonlinear system identification.The problem becomes more complex if the system is completely unknown but only the output time series is available.Blind recognition problem is the direct consequence of such a situation.The thesis concentrates on such problems.Capability of Artificial Neural Networks to approximate many nonlinear input-output maps makes it predominantly suitable for building a function for the identification of nonlinear systems,where only the time series is available.The literature is rich with a variety of algorithms to train the Neural Network model.A comprehensive study of the computation of the model parameters,using the different algorithms and the comparison among them to choose the best technique is still a demanding requirement from practical system designers,which is not available in a concise form in the literature.The thesis is thus an attempt to develop and evaluate some of the well known algorithms and propose some new techniques,in the context of Blind recognition of nonlinear systems.It also attempts to establish the relative merits and demerits of the different approaches.comprehensiveness is achieved in utilizing the benefits of well known evaluation techniques from statistics. The study concludes by providing the results of implementation of the currently available and modified versions and newly introduced techniques for nonlinear blind system modeling followed by a comparison of their performance.It is expected that,such comprehensive study and the comparison process can be of great relevance in many fields including chemical,electrical,biological,financial and weather data analysis.Further the results reported would be of immense help for practical system designers and analysts in selecting the most appropriate method based on the goodness of the model for the particular context.

Feature Extraction Methods Based on Linear Predictive Coding and Wavelet Packet Decomposition for Recognizing Spoken Words in Malayalam

Speech signals are one of the most important means of communication among the human beings. In this paper, a comparative study of two feature extraction techniques are carried out for recognizing speaker independent spoken isolated words. First one is a hybrid approach with Linear Predictive Coding (LPC) and Artificial Neural Networks (ANN) and the second method uses a combination of Wavelet Packet Decomposition (WPD) and Artificial Neural Networks. Voice signals are sampled directly from the microphone and then they are processed using these two techniques for extracting the features. Words from Malayalam, one of the four major Dravidian languages of southern India are chosen for recognition. Training, testing and pattern recognition are performed using Artificial Neural Networks. Back propagation method is used to train the ANN. The proposed method is implemented for 50 speakers uttering 20 isolated words each. Both the methods produce good recognition accuracy. But Wavelet Packet Decomposition is found to be more suitable for recognizing speech because of its multi-resolution characteristics and efficient time frequency localizations

Predicting sense of community and participation by applying machine learning to open government data

Community capacity is used to monitor socio-economic development. It is composed of a number of dimensions, which can be measured to understand the possible issues in the implementation of a policy or the outcome of a project targeting a community. Measuring community capacity dimensions is usually expensive and time consuming, requiring locally organised surveys. Therefore, we investigate a technique to estimate them by applying the Random Forests algorithm on secondary open government data. This research focuses on the prediction of measures for two dimensions: sense of community and participation. The most important variables for this prediction were determined. The variables included in the datasets used to train the predictive models complied with two criteria: nationwide availability; sufficiently fine-grained geographic breakdown, i.e. neighbourhood level. The models explained 77% of the sense of community measures and 63% of participation. Due to the low geographic detail of the outcome measures available, further research is required to apply the predictive models to a neighbourhood level. The variables that were found to be more determinant for prediction were only partially in agreement with the factors that, according to the social science literature consulted, are the most influential for sense of community and participation. This finding should be further investigated from a social science perspective, in order to be understood in depth.

Application of neural network to hybrid systems with binary inputs

Boolean input systems are in common used in the electric industry. Power supplies include such systems and the power converter represents these. For instance, in power electronics, the control variable are the switching ON and OFF of components as thyristors or transistors. The purpose of this paper is to use neural network (NN) to control continuous systems with Boolean inputs. This method is based on classification of system variations associated with input configurations. The classical supervised backpropagation algorithm is used to train the networks. The training of the artificial neural network and the control of Boolean input systems are presented. The design procedure of control systems is implemented on a nonlinear system. We apply those results to control an electrical system composed of an induction machine and its power converter.

Classification-based probabilistic modeling of texture transition for fast line search tracking and delineation

We introduce a classification-based approach to finding occluding texture boundaries. The classifier is composed of a set of weak learners, which operate on image intensity discriminative features that are defined on small patches and are fast to compute. A database that is designed to simulate digitized occluding contours of textured objects in natural images is used to train the weak learners. The trained classifier score is then used to obtain a probabilistic model for the presence of texture transitions, which can readily be used for line search texture boundary detection in the direction normal to an initial boundary estimate. This method is fast and therefore suitable for real-time and interactive applications. It works as a robust estimator, which requires a ribbon-like search region and can handle complex texture structures without requiring a large number of observations. We demonstrate results both in the context of interactive 2D delineation and of fast 3D tracking and compare its performance with other existing methods for line search boundary detection.

Prediction of Parkinson’s Disease tremor onset using a radial basis function neural network based on particle swarm optimization

Deep Brain Stimulation (DBS) has been successfully used throughout the world for the treatment of Parkinson's disease symptoms. To control abnormal spontaneous electrical activity in target brain areas DBS utilizes a continuous stimulation signal. This continuous power draw means that its implanted battery power source needs to be replaced every 18–24 months. To prolong the life span of the battery, a technique to accurately recognize and predict the onset of the Parkinson's disease tremors in human subjects and thus implement an on-demand stimulator is discussed here. The approach is to use a radial basis function neural network (RBFNN) based on particle swarm optimization (PSO) and principal component analysis (PCA) with Local Field Potential (LFP) data recorded via the stimulation electrodes to predict activity related to tremor onset. To test this approach, LFPs from the subthalamic nucleus (STN) obtained through deep brain electrodes implanted in a Parkinson patient are used to train the network. To validate the network's performance, electromyographic (EMG) signals from the patient's forearm are recorded in parallel with the LFPs to accurately determine occurrences of tremor, and these are compared to the performance of the network. It has been found that detection accuracies of up to 89% are possible. Performance comparisons have also been made between a conventional RBFNN and an RBFNN based on PSO which show a marginal decrease in performance but with notable reduction in computational overhead.

Reduction of invasive tests in chagasic patients with a modified self-organizing map

The applicability of AI methods to the Chagas' disease diagnosis is carried out by the use of Kohonen's self-organizing feature maps. Electrodiagnosis indicators calculated from ECG records are used as features in input vectors to train the network. Cross-validation results are used to modify the maps, providing an outstanding improvement to the interpretation of the resulting output. As a result, the map might be used to reduce the need for invasive explorations in chronic Chagas' disease.

Wavelet packets for lighting-effects determination

This paper presents a system to determine lighting effiects within face images. The theories of multivariate discriminant analysis and wavelet packets transform are utilised to develop the proposed system. An extensive set of face images of different poses, illuminated from different angles, are used to train the system. The performance of the proposed system is evaluated by conducting experiments on different test sets, and by comparing its results against those of some existing counterparts.

Study of polycyclic aromatic hydrocarbons emissions from burning of local wood species in Australia

Four kinds of Australian local wood species were burned in a domestic wood heater installed in a laboratory. The selected wood species include pine, red gum, yellow box and sugar gum, that are the most popular domestic fuel wood in Australia. Particulate matter emissions from burning of each load of wood were sampled isokinetically on filter media from the flue by standard stack emission sampling train. The particle laden filters then went for Gas Chromatography/ Mass Spectrometer (GC/MS) analysis to determine polycyclic aromatic hydrocarbons (PAHs) concentrations. The sampling was conducted under two different burning conditions – air inlet of the combustion chamber fully open and half open. Approximately 15 types of PAHs were detected. Emission factors were derived as microgram of PAHs /kg of wood burned. Total particulate emission factors were also obtained from gravimetric measurement before and after the sampling. PAH emission profiles for four species were generated from the results. Comparisons of emission factors have been conducted among different species of wood, as well as under different burning conditions, ie. fast burning and slowing burning. According to the derived emission factors, pine displayed the highest level of PAHs among the four species, followed by red gum and yellow box, whereas sugar gum showed the lowest level of PAHs. Emission factors were compared between each type of wood under two different burning conditions, the slow burning condition, which was air inlet half open, clearly showed higher PAH levels compared to the fast burning condition. Total PAH fractions on particulate matter were calculated and compared among wood types under two burning conditions. During the fast burning condition, red gum and pine have the higher percentage of PAH to total particulate matter emission than sugar gum and yellow box. When changed to slow burning, the PAH fraction on particulate matter are all increased with sugar gum having the largest increase.

Games within games: gambling and trickery in professional running

Professional running is an overtly gambling sport in which a clear objective is to maximise winnings from the bookmakers, which is achieved through a careful concealment of a runner’s ability. Professional runners seldom win more than one significant race. Races are deliberately lost until runners acquire a sufficiently lenient handicap to significantly improve their chances of winning a race of their choosing. Successes, kudos and identities in this sport are evaluated from the cleverness of the win, largely measured by the trainer’s effectiveness in executing a gambling coup. The money prizes given to runners may be significantly bettered from gambling winnings and making the most of these is the major emphasis for most successful runners and trainers. Drawing from an ethnographic study of this sport in Australia, the paper argues that the gambling strategies of runners and trainers can be understood as zero-sum games.

A self-learning system for identifying harmful network information

Network information identification is a “hot” topic currently. This paper designs a self-learning system using neural network algorithm for identifying the harmful network messages of both Chinese and English languages. The system segments the message into words and creates key word vector which characterizes the harmful network information. The BP algorithm is taken advantage of to train the neural network. The result of training and studying of the neural network can be applied onto many network applications based on message identification. The result of experiments demonstrates that our system has a high degree of accuracy.

Machine learning and small robot navigation

Developing successful navigation and mapping strategies is an essential part of autonomous robot research. However, hardware limitations often make for inaccurate systems. This project serves to investigate efficient alternatives to mapping an environment, by first creating a mobile robot, and then applying machine learning to the robot and controlling systems to increase the robustness of the robot system. My mapping system consists of a semi-autonomous robot drone in communication with a stationary Linux computer system. There are learning systems running on both the robot and the more powerful Linux system. The first stage of this project was devoted to designing and building an inexpensive robot. Utilizing my prior experience from independent studies in robotics, I designed a small mobile robot that was well suited for simple navigation and mapping research. When the major components of the robot base were designed, I began to implement my design. This involved physically constructing the base of the robot, as well as researching and acquiring components such as sensors. Implementing the more complex sensors became a time-consuming task, involving much research and assistance from a variety of sources. A concurrent stage of the project involved researching and experimenting with different types of machine learning systems. I finally settled on using neural networks as the machine learning system to incorporate into my project. Neural nets can be thought of as a structure of interconnected nodes, through which information filters. The type of neural net that I chose to use is a type that requires a known set of data that serves to train the net to produce the desired output. Neural nets are particularly well suited for use with robotic systems as they can handle cases that lie at the extreme edges of the training set, such as may be produced by "noisy" sensor data. Through experimenting with available neural net code, I became familiar with the code and its function, and modified it to be more generic and reusable for multiple applications of neural nets.

Real-time detection of parked vehicles from multiple image streams

We present a system to detect parked vehicles in a typical commercial parking complex using multiple streams of images captured through IP connected devices. Compared to traditional object detection techniques and machine learning methods, our approach is significantly faster in detection speed in the presence of multiple image streams. It is also capable of comparable accuracy when put to test against existing methods. And this is achieved without the need to train the system that machine learning methods require. Our approach uses a combination of psychological insights obtained from human detection and an algorithm replicating the outcomes of a SVM learner but without the noise that compromises accuracy in the normal learning process. The result is faster detection with comparable accuracy. Our experiments on images captured from a local test site shows very promising results for an implementation that is not only effective and low cost but also opens doors to new parking applications when combined with other technologies.