Florina Ungureanu

Eye tracking mouse for human computer interaction

In many cases, persons with neuro-locomotor disabilities have a good level of understanding and should use their eyes for communication. In this paper a reliable, mobile and low-cost system based on eye tracking mouse is presented. The eye movement is detected by a head mounted device and consequently the mouse cursor is moved on the screen. A click event denoting a pictogram selection is performed if the patient gazes a certain time the corresponding image on the screen.

Mobile embedded system for human computer communication in assistive technology

In this paper a new technology to communicate with people with neuro-locomotor disabilities using embedded systems and eye tracking approach is presented. The eye movement is detected by a special device and the voluntary eye blinking is correlated with a pictogram or keyword selection reflecting patients needs. The implemented eye tracking method uses image processing technique based on binarization algorithm.

Virtual reality based stroke recovery for upper limbs using leap motion

In this paper a virtual reality based stroke recovery system for upper limbs is described. The patient is immersed in the virtual environment through the use of an Oculus Rift device and interacts with the system by using Leap Motion as input device. The patient experience is enriched by providing haptic feedback when interacting with objects in the virtual environment. The recovery therapy relies on TRAVEE systems state-of-the-art paradigm of using augmented feedback during early recovery stages to create new recovery possibilities.

A Framework to Assess Cortical Activity of Visually Impaired Persons during Training with a Sensory Substitution Device

In this paper we describe a framework for assessing the cognitive and emotional activity of blind and visually impaired people in relation to the usage of a sensory substitution system (SSD). The overall objective is to aid the design and development of the SSD by understanding how the different choices in encoding and rendering the environmental information to the user, as well as training, affects the user perception and experience while using the system. To this end, a framework for EEG data acquisition and processing, tightly integrated with the SSD is proposed. The interpretation of the cortical activity of the visually impaired persons is performed in correlation with results from psychophysical tests, where accuracy and response times for various tasks are recorded. The framework allows for synchronized acquisition of EEG measurements and system/user events while training and testing with the SSD. The training and testing with the SSD is performed in simulated (virtual) environments. The hardware and software architecture of the framework is described. Preliminary results of the cortical activity analysis for blind users are presented to show case the exploitation of the proposed framework.

Energy-efficient compression algorithm for wireless transmission of physiological data

Wireless sensors networks is an active research topic. The sensor nodes are the main building blocks of these networks. There is a permanent concern for designing more and more efficient nodes in order to satisfy the demanding specifications of a sensor network. This paper introduces a new method for data compression that can be applied in wireless data communication of physiological measurements. The proposed method makes use of the Walsh-Hadamard transform by selecting a subset of the coefficients based on the information content of the signal. Its performance is compared with other algorithms using the following metrics: compression factor, reconstruction error and energy consumption.

Eye tracking performance improvement

In this paper, solutions to improve the stability and time response for head mounted eye tracking system are presented. The detection of eye pupil centre coordinates is affected by noise due to physiologic eye tremor, noisy images and the algorithm used for pupil detection. That leads to unstable point of gaze detection, affecting the eye tracking precision. Treating pupil centre coordinates as two separate signals, filters can be applied to improve the pupil detection stability. The point fixation can be more accurate detected by adding a snap to point technique fixation.

BCI and FES Based Therapy for Stroke Rehabilitation Using VR Facilities

In recent years, the assistive technologies and stroke rehabilitation methods have been empowered by the use of virtual reality environments and the facilities offered by brain computer interface systems and functional electrical stimulators. In this paper, a therapy system for stroke rehabilitation based on these revolutionary techniques is presented. Using a virtual reality Oculus Rift device, the proposed system ushers the patient in a virtual scenario where a virtual therapist coordinates the exercises aimed at restoring brain function. The electrical stimulator helps the patient to perform rehabilitation exercises and the brain computer interface system and an electrooculography device are used to determine if the exercises are executed properly. Laboratory tests on healthy people led to system validation from technical point of view. The clinical tests are in progress, but the preliminary results of the clinical tests have highlighted the good satisfaction degree of patients, the quick accommodation with the proposed therapy, and rapid progress for each user rehabilitation.

Neuromarketing and visual attention study using eye tracking techniques

This paper presents a study of human behavior related to different marketing stimuli based on pupillometry data. The consumers gaze points on products of interest are acquired with a remote eye tracker and the obtained heatmaps and timestamps are used for further analysis. An application for recording users visual attention on web pages and advertising slides is implemented. A sensor for electrodermal activity is also used in order to have an additional response regarding consumer arousal. The results reveal that the human emotions and visual attention are highly correlated with practical marketing applications.

Usability assessment of assistive technology for blind and visually impaired

In this paper we describe the usability assessment of a system designed to help blind and visually impaired people to navigate and perceive the environment. The proposed system is based on sensory substitution, remapping the vision stimuli into audio and haptic ones. The goal of this study is to aid the development of the sensory substitution device (SSD) by understanding how the different choices in encoding and rendering the environmental information affects the users perception and experience while using the system. The preliminary results are presented to show the usability and usefulness of the proposed system.

Multi-Objective Optimization of Feature Selection procedure for EEG signals classification

EEG is one of the biomarkers adequate for memory load assessment. Feature Selection (FS) routines for electroencephalogram (EEG) signals have been extensively studied in the past years. Current research is often based on machine learning algorithms. This paper investigates the impact of a new evolutionary approach to Multi-Objective Optimization (MOO) of FS routine for memory load classification using EEG. The main contributions of this study consist in: i) the analysis of candidate feature vectors regarding the accuracy of classification and parsimony; ii) the development of a multi-objective Genetic Algorithm (GA) with an improved selection of the preferred final solution, able to optimize the accuracy and the parsimony of any associated classifier. The feature vectors are classified using Support Vector Machine (SVM) and Random Forests (RF) classifiers. The suggested algorithms are verified on two benchmark datasets, optimizing feature selection for memory load level recognition using EEG data. The testing scenarios motivate the necessity of MOO for EEG signal feature selection. The superior performance of the proposed method recommends it as a preprocessing phase in EEG complex applications.