María Angélica González Arrieta

Augmented Reality: An Observational Study Considering the MuCy Model to Develop Communication Skills on Deaf Children

To develop different Communication Skills (CS) on deaf children (such as making signs, reading, writing or speaking) we propose a Sign Language Teaching Model (SLTM) called Multi-language Cycle for Sign Language Understanding (MuCy). Also, we conduct an observational study at the Association of Parents of Deaf Children of Salamanca (ASPAS) in order to measure the development of CS on deaf children by using a kit of Sign Language Pedagogical Materials (SLPMs), as well as the use of Augmented Reality (AR) as complementary tools for teaching Sign Language (SL) within a Collaborative Learning Environment with Mixed-Reality (CLEMR).

Neural Networks through Shared Maps in Mobile Devices

We introduce a hybrid system composed of a convolutional neural network and a discrete graphical model for image recognition. This system improves upon traditional sliding window techniques for analysis of an image larger than the training data by effectively processing the full input scene through the neural network in less time. The final result is then inferred from the neural network output through energy minimization to reach a more precize localization than what traditional maximum value class comparisons yield. These results are apt for applying this process in a mobile device for real time image recognition.

Automatic Prediction of Poisonous Mushrooms by Connectionist Systems

The research offers a quite simple view of methods to classify edible and poisonous mushrooms. In fact, we are looking for not only classification methods but also for an application which supports experts’ decisions. To achieve our aim, we will study different structures of neural nets and learning algorithms, and select the best one, according to the test results.

Texture Classification with Neural Networks

Texture classification poses a well known difficulty within computer vision systems. This paper reviews a method for image segmentation based on the classification of textures using artificial neural networks. The supervised machine learning system developed here is able to recognize and distinguish among multiple feature regions within one or more photographs, where areas of interest are characterized by the various patterns of color and shape they exhibit. The use of an enhancement filter to reduce sensitivity to illumination and orientation changes in images is explored, as well as various post-processing techniques to improve the classification results based on context grouping. Various applications of the system are examined, including the geographical segmentation of satellite images and a brief overview of the model’s performance when employed on a real time video stream.

Ear Biometrics: A Small Look at the Process of Ear Recognition

This document provides an approach to biometrics analysis which consists in the location and identification of ears in real time. Ear features, which is a stable biometric approach that does not vary with age, have been used for many years in the forensic science of recognition. The ear has all the properties that a biometric trait should have, i.e. uniqueness, permanence, universality and collectability. Because it is a field of study with potential growth, in this paper, we summarize some of the approaches to the detection and recognition in existing 2D images in order to provide a perspective on the possible future research and the develop of a practical application of some of these methodologies to create finally a functional application for identification and recognition of individuals from an image of the ear, the above in the context of intelligent surveillance and criminal identification, one of the most important areas in the processes of identification.

Neural Networks Using Hausdorff Distance, SURF and Fisher Algorithms for Ear Recognition

The purpose of this paper is to offer an approach in the biometrics analysis field, using ears to recognize people. This study uses Hausdorff distance as a preprocessing stage adding sturdiness to increase the performance filtering for the subjects to use for testing stage of the neural network. Then, the system computes Speeded Up Robust Features (SURF) and Fisher Linear Discriminant Analysis (LDA) as an input of two neural networks to detect and recognize a person by the patterns of its ear. To show the applied theory in the experimental results; it also includes an application developed with Microsoft .net. The investigation which enhances the ear recognition process showed robustness through the integration of Hausdorff, LDA and SURF in neural networks.

Ear Recognition with Neural Networks Based on Fisher and Surf Algorithms

This paper offers an approach to biometric analysis using ears for recognition. The ear has all the assets that a biometric trait should possess. Because it is a study field in potential growth, this research offers an approach using Speeded Up Robust Features (SURF) and Fisher Linear Discriminant Analysis (LDA) as an input of two neural networks with the purpose to detect and recognize a person by the patterns of its ear. It also includes the development of an application with .net to show experimental results of the applied theory. In the preprocessing task, the system adds sturdiness using Hausdorff distance to increase the performance filtering for the subjects to use in the testing stage of the neural network. To perform this study, we worked with the help of Avilas police school (Spain), where we built a database with approximately 300 ears. The investigation results shown that the integration of LDA and SURF in neural networks can improve the ear recognition process and provide robustness in changes of illumination and perception.

Hybrid System for Mobile Image Recognition through Convolutional Neural Networks and Discrete Graphical Models

A system is presented which combines deep neural networks with discrete inference techniques for the successful recognition of an image. The system presented builds upon the classical sliding window method but applied in parallel over an entire input image. The result is discretized by treating each classified window as a node in a markov random field and applying a minimization of its associated energy levels. Two important benefits are observed with this system: a gain in performance by virtue of the systems parallel nature, and an improvement in the localization precision due to the inherent connectivity between classified windows.

Shared Map Convolutional Neural Networks for Real-Time Mobile Image Recognition

We present a technique for improving the speed of a convolutional neural network applied to large input images through the optimization of the sliding window approach. Meaningful performance gains and memory bandwidth reduction can be obtained by processing images in this manner, factors which play a crucial role in the deployment of deep neural networks within mobile devices.

A Brief Approach to the Ear Recognition Process

This paper offers an approach to biometric analysis using ears for recognition. The ear has all the assets that a biometric trait should possess. Because it is a study field in potential growth, this study offers an approach using SURF features as an input of a neural network with the purpose to detect and recognize a person by the patterns of its ear, also includes, the development of an application with .net to show experimental results of the theory applied. Ear characteristics, which are a unchanging biometric approach that does not vary with age, have been used for several years in the forensic science of recognition, thats why the research gets important value in the present. To perform this study, we worked with the help of Police School of Avila, Province of Spain, we have built a database with approximately 300 ears.