Schubert R. Carvalho

Experimenting on the cognitive walkthrough with users

This paper presents a case study aiming to investigate which variant of the Think-Aloud Protocol (i.e., the Concurrent Think-Aloud and the Retrospective Think-Aloud) better integrates with the Cognitive Walkthrough with Users. To this end we performed a case study that involved twelve users and one usability evaluator. Usability problems uncovered by each method were evaluated to help us understand the strengths and weaknesses of the studied usability testing methods. The results suggest that 1) the Cognitive Walkthrough with Users integrates equally well with both the Think-Aloud Protocol variants; 2) the Retrospective Think-Aloud find more usability problems and 3) the Concurrent Think-Aloud is slightly faster to perform and was more cost effective. However, this is only one case study, and further research is needed to verify if the results are actually statistically significant.

Plant Classification from Leaf Textures

This work describes a methodology for plant classification based on the analysis of leaf textures by combining a multi-resolution technique, such as the two-dimensional (2D) Discrete Wavelet Transform (2D-DWT), statistical models and Gray-Level Co-occurrence Matrices (GLCM) in which some invariance (e.g. rotation and scale) are achieved. As a second step, an Artificial Neural Network (ANN) model is trained for automatic classifying plant species. The proposed approach was tested on the Flavia database. An overall classification accuracy of 91.85% was achieved which demonstrates that plants can be reliably classified using texture samples extracted from leaf tissues.

Exploring the Use of Leaf Shape Frequencies for Plant Classification

Plant identification and classification play an important role in ecology, but the manual process is cumbersome even for experimented taxonomists. Technological advances allows the development of strategies to make these tasks easily and faster. In this context, this paper describes a methodology for plant identification and classification based on leaf shapes, that explores the discriminative power of the contour-centroid distance in the Fourier frequency domain in which some invariance (e.g. Rotation and scale) are guaranteed. In addition, it is also investigated the influence of feature selection techniques regarding classification accuracy. Our results show that by combining a set of features vectors - in the principal components space - and a feed forward neural network, an accuracy of 97.45% was achieved.

A deep-learning-based approach for automated wagon component inspection

Inspecting objects in the industry aims to guarantee product quality allowing problems to be corrected and damaged products to be discarded. Inspection is also widely used in railway maintenance, where wagon components need to be checked due to efficiency and safety concerns. In some organizations, hundreds of wagons are inspected visually by a human inspector, which leads to quality issues and safety risks for the inspectors. This paper describes a wagon component inspection approach using Deep Learning techniques to detect a particular damaged component: the shear pad. We compared our approach for convolutional neural networks with the state of art classification methods to distinguish among three shear pads conditions: absent, damaged, and undamaged shear pad. Our results are very encouraging showing empirical evidence that our approach has better performance than other classification techniques.

Anticipatory EEG Signals for Detecting and Classifying Game Interaction Onset

In this study, we propose to investigate the presence of anticipatory patterns in electroencephalography (EEG) signals while driving a virtual car to determine two specific actions (1) turn left and (2) turn right, a few milliseconds before such actions take place. Our results show the feasibility of using anticipatory brain signals for detecting and classifying game interaction before it happens.