Daniela Iacoviello

Robust real time eye tracking for computer interface for disabled people

Gaze is a natural input for a Human Computer Interface (HCI) for disabled people, who have of course an acute need for a communication system. An efficient eye tracking procedure is presented providing a non-invasive method for real time detection of a subject eyes in a sequence of frames captured by low cost equipment. The procedure can be easily adapted to any subject and is adequately insensitive to changing of the illumination. The eye identification is performed on a piece-wise constant approximation of the frames. It is based on a discrete level set formulation of the variational approach to the optimal segmentation problem. This yields a simplified version of the original data retaining all the information relevant to the application. Tracking is obtained by a fast update of the optimal segmentation between successive frames. No eye movement model is required being the procedure fast enough to obtain the current frame segmentation as one step update from the previous frame segmentation.

A real-time classification algorithm for EEG-based BCI driven by self-induced emotions

BACKGROUND AND OBJECTIVE The aim of this paper is to provide an efficient, parametric, general, and completely automatic real time classification method of electroencephalography (EEG) signals obtained from self-induced emotions. The particular characteristics of the considered low-amplitude signals (a self-induced emotion produces a signal whose amplitude is about 15% of a really experienced emotion) require exploring and adapting strategies like the Wavelet Transform, the Principal Component Analysis (PCA) and the Support Vector Machine (SVM) for signal processing, analysis and classification. Moreover, the method is thought to be used in a multi-emotions based Brain Computer Interface (BCI) and, for this reason, an ad hoc shrewdness is assumed. METHOD The peculiarity of the brain activation requires ad-hoc signal processing by wavelet decomposition, and the definition of a set of features for signal characterization in order to discriminate different self-induced emotions. The proposed method is a two stages algorithm, completely parameterized, aiming at a multi-class classification and may be considered in the framework of machine learning. The first stage, the calibration, is off-line and is devoted at the signal processing, the determination of the features and at the training of a classifier. The second stage, the real-time one, is the test on new data. The PCA theory is applied to avoid redundancy in the set of features whereas the classification of the selected features, and therefore of the signals, is obtained by the SVM. RESULTS Some experimental tests have been conducted on EEG signals proposing a binary BCI, based on the self-induced disgust produced by remembering an unpleasant odor. Since in literature it has been shown that this emotion mainly involves the right hemisphere and in particular the T8 channel, the classification procedure is tested by using just T8, though the average accuracy is calculated and reported also for the whole set of the measured channels. CONCLUSIONS The obtained classification results are encouraging with percentage of success that is, in the average for the whole set of the examined subjects, above 90%. An ongoing work is the application of the proposed procedure to map a large set of emotions with EEG and to establish the EEG headset with the minimal number of channels to allow the recognition of a significant range of emotions both in the field of affective computing and in the development of auxiliary communication tools for subjects affected by severe disabilities.

Coupling image processing and stress analysis for damage identification in a human premolar tooth

Non-carious cervical lesions are characterized by the loss of dental hard tissue at the cement-enamel junction (CEJ). Exceeding stresses are therefore generated in the cervical region of the tooth that cause disruption of the bonds between the hydroxyapatite crystals, leading to crack formation and eventual loss of enamel and the underlying dentine. Damage identification was performed by image analysis techniques and allowed to quantitatively assess changes in teeth. A computerized two-step procedure was generated and applied to the first left maxillary human premolar. In the first step, dental images were digitally processed by a segmentation method in order to identify the damage. The considered morphological properties were the enamel thickness and total area, the number of fragments in which the enamel is chipped. The information retrieved by the data processing of the section images allowed to orient the stress investigation toward selected portions of the tooth. In the second step, a three-dimensional finite element model based on CT images of both the tooth and the periodontal ligament was employed to compare the changes occurring in the stress distributions in normal occlusion and malocclusion. The stress states were analyzed exclusively in the critical zones designated in the first step. The risk of failure at the CEJ and of crack initiation at the dentin-enamel junction through the quantification of first and third principal stresses, von Mises stress, and normal and tangential stresses, were also estimated.

Optimal bone density distributions

In this paper a control and optimization procedure for bone remodeling simulations was adopted to study the effect of the osteocyte influence range on the predicted density distribution. In order to reach this goal, the osteocyte network regulating bone remodeling process in a 2-D bone sample was numerically simulated. The assumed proportional-integral-derivative (PID) bone remodeling rule was related to the error signal between the strain energy density and a selected target. Furthermore the control parameters and the target were optimally determined minimizing a suitable cost index: the goal was to minimize the final mass and the energy thus maximizing the stiffness. The continuum model results show that the developed and adapted trabecular structure was consistent with the applied loads and only depended on the external forces, the value of the cost index, the maximum attainable elastic modulus value (hence, the maximum density value) and the value of the energy target. The remodeling phenomenon determined the number and thickness of the trabeculae which are formed from a uniform distribution of mass density in the considered domain; this number and these thicknesses are controlled by the values assigned to the parameters of the model. In particular, the osteocyte decay distance (D) of the influence range affected the trabecular patterns formation, showing an important effect in the adaptive capacity of the optimization numerical model.

Overall design and implementation of the virtual glove

Post-stroke patients and people suffering from hand diseases often need rehabilitation therapy. The recovery of original skills, when possible, is closely related to the frequency, quality, and duration of rehabilitative therapy. Rehabilitation gloves are tools used both to facilitate rehabilitation and to control improvements by an evaluation system. Mechanical gloves have high cost, are often cumbersome, are not re-usable and, hence, not usable with the healthy hand to collect patient-specific hand mobility information to which rehabilitation should tend. The approach we propose is the virtual glove, a system that, unlike tools based on mechanical haptic interfaces, uses a set of video cameras surrounding the patient hand to collect a set of synchronized videos used to track hand movements. The hand tracking is associated with a numerical hand model that is used to calculate physical, geometrical and mechanical parameters, and to implement some boundary constraints such as joint dimensions, shape, joint angles, and so on. Besides being accurate, the proposed system is aimed to be low cost, not bulky (touch-less), easy to use, and re-usable. Previous works described the virtual glove general concepts, the hand model, and its characterization including system calibration strategy. The present paper provides the virtual glove overall design, both in real-time and in off-line modalities. In particular, the real-time modality is described and implemented and a marker-based hand tracking algorithm, including a marker positioning, coloring, labeling, detection and classification strategy, is presented for the off-line modality. Moreover, model based hand tracking experimental measurements are reported, discussed and compared with the corresponding poses of the real hand. An error estimation strategy is also presented and used for the collected measurements. System limitations and future work for system improvement are also discussed.

Optimal segmentation of pupillometric images for estimating pupil shape parameters

The problem of determining the pupil morphological parameters from pupillometric data is considered. These characteristics are of great interest for non-invasive early diagnosis of the central nervous system response to environmental stimuli of different nature, in subjects suffering some typical diseases such as diabetes, Alzheimer disease, schizophrenia, drug and alcohol addiction. Pupil geometrical features such as diameter, area, centroid coordinates, are estimated by a procedure based on an image segmentation algorithm. It exploits the level set formulation of the variational problem related to the segmentation. A discrete set up of this problem that admits a unique optimal solution is proposed: an arbitrary initial curve is evolved towards the optimal segmentation boundary by a difference equation; therefore no numerical approximation schemes are needed, as required in the equivalent continuum formulation usually adopted in the relevant literature.

A discrete level set approach to image segmentation

Models and algorithms in image processing are usually defined in the continuum and then applied to discrete data, that is the signal samples over a lattice. In particular, the set up in the continuum of the segmentation problem allows a fine formulation basically through either a variational approach or a moving interfaces approach. In any case, the image segmentation is obtained as the steady-state solution of a nonlinear PDE. Nevertheless the application to real data requires discretization schemes where some of the basic image geometric features have a loose meaning. In this paper, a discrete version of the level set formulation of a modified Mumford and Shah energy functional is investigated, and the optimal image segmentation is directly obtained through a nonlinear finite difference equation. The typical characteristics of a segmentation, such as its component domains area and its boundary length, are all defined in the discrete context thus obtaining a more realistic description of the available data. The existence and uniqueness of the optimal solution is proved in the class of piece wise constant functions, but with no restrictions on the nature of the segmentation boundary multiple points. The proposed algorithm compared to a standard segmentation procedure in the continuum generally provides a more accurate segmentation, with a much lower computational cost.

Parametric characterization of the form of the human pupil from blurred noisy images

The fluctuation of the human pupil is an important parameter in order to make non-invasive diagnosis of many different diseases and in several clinical applications. The relevant measurement device, the pupillometer, consists in a CCD camera, which shoots the pupil. We suppose that the measured image is blurred by a Gaussian kernel and corrupted by an additive white noise; moreover an elliptic shape for the pupil is assumed. We here present the extension of a multiscale approach for edge detection to identify some parameters of the pupil: the location of its centre, the length of the semi-axes and the orientation of the corresponding ellipse. The chosen method requires knowledge about the degradation parameters of the assumed model; so we first present a simple but efficient method to determine such quantities for the measured image. Then we apply the edge detection procedure to identify points close to the pupil edge, within a chosen probability. Finally we find the optimal ellipse fitting a suitable subset of the previously detected edge points. Results are presented, with comparisons to other approaches for edge finding.

Modeling for edge detection problems in blurred noisy images

The aim of this paper is to provide a theoretical set up and a mathematical model for the problem of image reconstruction. The original image belongs to a family of two-dimensional (2-D) possibly discontinuous functions, but is blurred by a Gaussian point spread function introduced by the measurement device. In addition, the blurred image is corrupted by an additive noise. We propose a preprocessing of data which enhances the contribution of the signal discontinuous component over that one of the regular part, while damping down the effect of noise. In particular we suggest to convolute data with a kernel defined as the second order derivative of a Gaussian spread function. Finally, the image reconstruction is embedded in an optimal problem framework. Now convexity and compactness properties for the admissible set play a fundamental role. We provide an instance of a class of admissible sets which is relevant from an application point of view while featuring the desired properties.

A low-cost real time virtual system for postural stability assessment at home

BACKGROUND AND OBJECTIVE The degeneration of the balance control system in the elderly and in many pathologies requires measuring the equilibrium conditions very often. In clinical practice, equilibrium control is commonly evaluated by using a force platform (stabilometric platform) in a clinical environment. In this paper, we demonstrate how a simple movement analysis system, based on a 3D video camera and a 3D real time model reconstruction of the human body, can be used to collect information usually recorded by a physical stabilometric platform. METHODS The algorithm used to reconstruct the human body model as a set of spheres is described and discussed. Moreover, experimental measurements and comparisons with data collected by a physical stabilometric platform are also reported. The measurements were collected on a set of 6 healthy subjects to whom a change in equilibrium condition was stimulated by performing an equilibrium task. RESULTS The experimental results showed that more than 95% of data collected by the proposed method were not significantly different from those collected by the classic platform, thus confirming the usefulness of the proposed system. CONCLUSIONS The proposed virtual balance assessment system can be implemented at low cost (about 500