Víctor Uc-Cetina

Usability evaluation of an augmented reality system for teaching Euclidean vectors

Abstract Augmented reality (AR) is one of the emerging technologies that has demonstrated to be an efficient technological tool to enhance learning techniques. In this paper, we describe the development and evaluation of an AR system for teaching Euclidean vectors in physics and mathematics. The goal of this pedagogical tool is to facilitate user’s understanding of physical concepts, such as magnitude, direction and orientation, together with basic vector-related operations like addition, subtraction and cross product. The result of the system usability scale showed our system’s usability and learnability. The system merges a real-world scenario with virtual elements controlled with a practical body-interactive interface.

Composition of Web Services Using Markov Decision Processes and Dynamic Programming

We propose a Markov decision process model for solving the Web service composition (WSC) problem. Iterative policy evaluation, value iteration, and policy iteration algorithms are used to experimentally validate our approach, with artificial and real data. The experimental results show the reliability of the model and the methods employed, with policy iteration being the best one in terms of the minimum number of iterations needed to estimate an optimal policy, with the highest Quality of Service attributes. Our experimental work shows how the solution of a WSC problem involving a set of 100,000 individual Web services and where a valid composition requiring the selection of 1,000 services from the available set can be computed in the worst case in less than 200 seconds, using an Intel Core i5 computer with 6 GB RAM. Moreover, a real WSC problem involving only 7 individual Web services requires less than 0.08 seconds, using the same computational power. Finally, a comparison with two popular reinforcement learning algorithms, sarsa and Q-learning, shows that these algorithms require one or two orders of magnitude and more time than policy iteration, iterative policy evaluation, and value iteration to handle WSC problems of the same complexity.

Chagas Parasite Detection in Blood Images Using AdaBoost

The Chagas disease is a potentially life-threatening illness caused by the protozoan parasite, Trypanosoma cruzi. Visual detection of such parasite through microscopic inspection is a tedious and time-consuming task. In this paper, we provide an AdaBoost learning solution to the task of Chagas parasite detection in blood images. We give details of the algorithm and our experimental setup. With this method, we get 100% and 93.25% of sensitivity and specificity, respectively. A ROC comparison with the method most commonly used for the detection of malaria parasites based on support vector machines (SVM) is also provided. Our experimental work shows mainly two things: (1) Chagas parasites can be detected automatically using machine learning methods with high accuracy and (2) AdaBoost + SVM provides better overall detection performance than AdaBoost or SVMs alone. Such results are the best ones known so far for the problem of automatic detection of Chagas parasites through the use of machine learning, computer vision, and image processing methods.

An automatic algorithm for the detection of Trypanosoma cruzi parasites in blood sample images

Chagas disease is a tropical parasitic disease caused by the flagellate protozoan Trypanosoma cruzi (T. cruzi) and currently affecting large portions of the Americas. One of the standard laboratory methods to determine the presence of the parasite is by direct visualization in blood smears stained with some colorant. This method is time-consuming, requires trained microscopists and is prone to human mistakes. In this article we propose a novel algorithm for the automatic detection of T. cruzi parasites, in microscope digital images obtained from peripheral blood smears treated with Wrights stain. Our algorithm achieved a sensitivity of 0.98 and specificity of 0.85 when evaluated against a dataset of 120 test images. Experimental results show the versatility of the method for parasitemia determination.

Dynamic web services selection using a hidden Markov model

In a service oriented architecture based on web services exists the possibility of failures occurring at the time a transaction between web services runs. These failures are undesired because they reduce the systems performance. Self-healing systems are based on the model of the human body to restore it from an unhealthy state to a healthy one. These self-healing systems represent a good option for handling failures in a system based on web services. However, self-healing systems could improve their performance by adding a mechanism than selects the most suitable web services to perform certain functions. Moreover the diagnosis module of the self healing system would benefit by reducing the failures situations caused by anomalous web services. In this paper we propose to reduce the number of systems failures by employing a hidden Markov model that assist in the selection of web services through the use a QoS-based model. By reducing the number of failures this mechanism would support the diagnosis module of a self-healing system.

Glove-based sign language recognition solution to assist communication for deaf users

This manuscript presents the research and development of a software that help deaf-mute communication by identifying the position of the fingers of the hand with 5DT gloves. The sign language is adopted by nearly all people with hearing deficiency, making it their main form of communication, but this communication is only successfully achieved if all the participants of the conversation are familiar with the sign language. The goal is to be able to translate hand signs into words and phrases with the possibility to send audio signals to allow deaf-mute users to communicate to people not familiar with the sign language. The recognition of hand gestures is accomplished using a neural network tested using five different training algorithms. A cross-validation experiment is provided to illustrate the robustness of our methods.

Towards an automatic counter of lunar craters

Quantification of impact craters on planetary surfaces is relevant to understand the geological history of the planet. In order to automatize quantification of lunar craters in digital images, the first step is to develop a computational tool capable of classifying a subwindow of pixels into two possible outputs: crater / non-crater. In this paper, we provide preliminary experimental results using an adaptive boosting algorithm to train a binary classifier for lunar crater identification. Using 30 weak classifiers we obtain 0.925 and 0.94 of sensitivity and specificity, respectively.

A novel reinforcement learning architecture for continuous state and action spaces

We introduce a reinforcement learning architecture designed for problems with an infinite number of states, where each state can be seen as a vector of real numbers and with a finite number of actions, where each action requires a vector of real numbers as parameters. The main objective of this architecture is to distribute in two actors the work required to learn the final policy. One actor decideswhat actionmust be performed;meanwhile, a second actor determines the right parameters for the selected action.We tested our architecture and one algorithmbased on it solving the robot dribbling problem, a challenging robot control problem taken from the RoboCup competitions. Our experimental work with three different function approximators provides enough evidence to prove that the proposed architecture can be used to implement fast, robust, and reliable reinforcement learning algorithms.We introduce a reinforcement learning architecture designed for problems with an infinite number of states, where each state can be seen as a vector of real numbers and with a finite number of actions, where each action requires a vector of real numbers as parameters. The main objective of this architecture is to distribute in two actors the work required to learn the final policy. One actor decideswhat actionmust be performed;meanwhile, a second actor determines the right parameters for the selected action. We tested our architecture and one algorithmbased on it solving the robot dribbling problem, a challenging robot control problem taken from the RoboCup competitions. Our experimental work with three different function approximators provides enough evidence to prove that the proposed architecture can be used to implement fast, robust, and reliable reinforcement learning algorithms.