Six out of eight research projects submitted by TICKET Lab's researchers have been accepted:

Preserving Multimedia Data Quality and Privacy (Dr. Chady Abou Jaoudé):

In today's world, social media is invading every person’s daily life. Millions of people have at least one account in each and every social media application; they share thoughts, photos, videos and all kinds of data revealing their personal information. The huge number of data shared on social media raises some major privacy concerns, as the data collected is often privacy-sensitive. In this work, we intend to provide a method to allow users to protect their privacy by themselves. To do so, we will define a framework that processes a stream of multimedia data using a well-established security rules to protect the entities appearing in the images. Moreover, we will compute a trade-off between data quality and data privacy by quantifying the utility of a protected image to which several masking functions have been applied on the basis of two main characteristics: the affected area of protection and the semantic coherence of the protected item.


On the Evaluation of Security and Privacy in Smart Cities (Dr. Béchara Al Bouna):

In today's world, we find ourselves surrounded by many IoT-based cyber-physical systems that silently track our activities and collect sensitive information about us. Among the most prominent examples, we cite smart environments (e.g., smart homes and cities), quantified self-technologies, smart energy meters, etc. While such systems promise to ease our lives, they raise major privacy concerns for their users, as the data collected is often privacy-sensitive, such as location of individuals, patients' vital signs. In this work, we will investigate the methodology to empower users to protect their privacy by themselves. That is, users should be able, before sharing a private data item (or a combination of data items) with a data consumer, to:

• Understand the privacy risks involved in that sharing;

• Assess the value of the data to be shared, based on the identified privacy risks, and compare it to the potential benefits generated by the sharing;

• Negotiate with data consumers to attain a (trade-off) data sharing decision satisfying both parties when conflicts happen;

• Control the data release by applying the necessary data modification techniques (e.g., anonymization, data perturbation, modification, etc.) to implement the attained sharing decision.


Mašriq Traditional Modal Monodies Encoding in the Music Encoding Initiative (MEI) (Dr. Talar Atéchian):

Music encoding is a representation of music sheets or music transcriptions to machine-readable structure. In this project, traditional Mašriq’s modal monodies transcriptions are encoded using Music Encoding Initiative (MEI) standard. In addition to the fundamental music elements encoded in the transcriptions, Modal Semiotic Theory [1] is used, providing a new encoding dimension. Thus, a generative grammar is proposed in the Theory to enrich the encoding process and to provide a detailed interpretation of the analyzed modal monodies. A semi-automated algorithm is implemented to encode in MEI standard modal monodies transcriptions. The prototype returns as a result, encoded music transcriptions accompanied by a detailed semiotic analysis. Experimentations are conducted, and still in progress, to evaluate the performance of the proposed algorithm, and the quality of the returned analysis. As a second phase of the project, we aim to develop a fully automated algorithm for the modal monodies encoding. Mathematical models and machine learning algorithms will be implemented. The fully automated open source prototype will be developed and the performance the accuracy of the returned results will be evaluated by musicologists.


Data Mining and Machine Learning to Extract Important Markers to Improve Medical Monitoring Platforms (Dr. Georges Badr):

Chronic diseases are responsible for rising health spending in most developed countries. Early detection, prevention and treatment of long-term complications of chronic diseases should help limit spending and provide a better quality of life for patients. Patient follow-up monitors their medical parameters and interprets them to detect at-risk situations early and assist caregivers in their diagnosis. The reasoning used here is based on the physician's decision-making process, with both theoretical and empirical knowledge constructed in the course of his or her experience. Early detection of heart failure requires the supervision of various vital signs and lifestyle. These elements, consolidated with the patient's profile, allow the detection of cardiac abnormalities and the prevention of risk situations. Data mining technologies produce new knowledge based on the search for motives characterizing the symptoms from the first occurrence. This characterization, which takes into account the evolution of all the vital signs associated with habits and lifestyle, will allow us to formalize detection.


Analyse de stress quotidien chez les non-voyants(Dr. Youssef Bou Issa):

Ce projet s’inscrit dans le cadre des villes/maisons intelligentes dans le but d’améliorer la qualité de vie des personnes non-voyantes. En effet, notre objectif est d’étudier le stress des personnes non-voyantes dans leur vie quotidienne et ceci en analysant leurs activités dans les périodes de la journée. Cette analyse se réalise suite à l’acquisition, traitement et fusion des données issues des réseaux de capteurs corporels. Par la suite l’objectif principal est de réaliser une plateforme permettant d’acquérir les données à partir de capteurs corporels non-invasifs. Ces capteurs permettent des récupérer en temps réel des données liées à l’intensité de l’activité physique et cérébrale d’une personne. Ces données nous permettront d’élaborer des indicateurs de stress et les facteurs influençant. En effet, ce travail constitue une suite de nos travaux précédents concernant l’analyse de stress chez des employés voyants en utilisant les réseaux de capteurs. Par la suite, nous pourrons réaliser une étude comparative du stress chez les voyants et chez les non-voyants.


Design of Reconfigurable Reflectarray Antenna Based on Varicap Diodes (Dr. Tony Makdissy):

Reflectarray forms a new family of antenna inheriting its properties from reflector antennas and from phased array antennas. A reflectarray consists of a primary source (typically a horn antenna) illuminating a flat panel of radiating elements called “phase-shifting cells”. The shape of the resulting radiation pattern is obtained by controlling the phase of the wave reflected by these cells.  The investigation on active techniques to control the reflected phase (using control elements such as MEMS switches, PIN diodes or varicap diodes) opens the field of reconfigurable reflectarrays. The main challenge while designing a reconfigurable reflectarray is to design a phase-shifting cell loaded with a limited number of control elements. In addition, the cell must be able to provide uniformly distributed phase states (or continuous variation of the phase) over 360 degrees. Finally, the variation of the different phase states with the frequency must be linear and parallel with low frequency dispersion to ensure a wideband behavior. This project aims to design two phase-shifting cells loaded with varicap diodes for single and dual linear polarizations reflectarray. The designed cells will be tested on real reflectarrays in order to demonstrate the reconfigurability of the radiation pattern.