Multimedia Technologies in Biometrics and Information Security Applications

Catalogue of disciplines

Multimedia technologies are currently a fast-growing field. On the one hand, digital sound and image technology has become ubiquitous in all areas of society and industry, and the transition to digital technology is almost complete. On the other hand, the use of modern artificial intelligence (AI) techniques in this field has been increasingly observed in recent years, with already impressive results in processing speech and audio or image data. In this respect, several important directions can be highlighted in which multimedia technologies are expected to play an important role in the coming period:

• Biometric technologies (methods of uniquely recognising people on the basis of one or more intrinsic physical or behavioural traits) have evolved greatly over the last decade. Nowadays, due to their widespread use as identification, security and protection features or their use in commercial applications, biometric-based technologies need to be studied. For such applications, training in this area is necessary both for development and implementation, and for understanding the market and the requirements of users and manufacturers of such systems.
• Intelligent voice communication interfaces are already commonplace in our modern world, there is a clear requirement to make it as simple, natural and efficient as possible to interact with robots and computer systems (fixed or portable) or to access various services and information remotely. The design and development of such systems is today increasingly based on new artificial intelligence techniques used in speech and natural language processing, which are the only ones capable of leading to evolved dialogue systems.
• As a science in its own right, forensics appeared as an objective necessity, determined by the fact that the means available to the justice system to counter criminal acts had become insufficient. The scale of crime and the diversification of the methods and means used by criminals have made it necessary to improve judicial activity and give a modern character to the fight against crime. The judicial authorities are often confronted with problems they cannot solve alone, and this requires the expertise of specialists in scientific fields. More and more countries are introducing audio-video recordings as evidence, and a thorough knowledge of the main techniques for tampering or manipulating audio-video signals, as well as methods for forensically examining and verifying the authenticity of files containing audio and/or video information, is vital in the field of forensic science.

Alongside multimedia technologies, information security is also a fundamental requirement of today's society; cyber security is, in fact, the fastest growing area in the information and computer technology (ICT) industry. Information security involves protocols, technologies, systems, tools and techniques to secure and stop malicious attacks, attacks that can lead to the loss or theft of sometimes critical information from institutions or companies. As our ability to collect, process and distribute information continues to grow, the demand for securing computers, computer networks and IoT devices is growing even faster.

The BIOSINF Master's Research Programme is primarily aimed at students who have completed their undergraduate studies in the fundamental field of engineering sciences (Electronic Engineering, Telecommunications and Information Technology, Computer and Information Technology, Applied Engineering Sciences, Electrical Engineering), but can also be attended by students from related faculties, such as those in the basic sciences (mathematics, physics, computer science), interested in applications of information technology in cutting-edge areas such as multimedia technologies, artificial intelligence, biometrics and information security. This preliminary definition of the main target group does not exclude its extension, at the time of admission to the BIOSINF Master programme, with students graduating from other undergraduate studies.

The BIOSINF Master programme, which started in the 2011-2012 academic year, has two main objectives:

• Acquire theoretical and practical knowledge in the field of digital analysis and processing of audio, speech and video images and sequences, in particular using modern techniques based on artificial intelligence (classical machine learning methods and deep neural networks), with applications in biometric authentication, human-machine interaction, forensic analysis and forensics of audio-video recordings (in particular speech forensics) and implementation of resource-constrained multimedia systems.
• Acquire the essential knowledge needed to design and operate information systems securely, gain expertise in investigating cybersecurity incidents in computer networks or IoT devices, and develop security policies needed to prevent risks and threats to digital infrastructures.

• Ability to understand, select, design and implement specific algorithms for the analysis and digital processing of voice, image and audio-video signals.
• Understanding how information is represented in advanced signal processing systems and knowledge of modern data classification techniques based on artificial intelligence (machine learning and deep neural networks).
• Familiarity with the field of behavioural biometrics, especially speaker recognition, and the mathematical tools used.
• Practical implementation of voice chat-bot systems based on speech technology methods.
• Ability to use interdisciplinary (including legal) knowledge and to use conceptual tools specific to the 'forensic science' branch.
• Ability to understand, design and perform analysis on audio-video recordings in order to detect traces of previous editing/manipulation, detect and identify the person speaking or appearing in the footage, and authenticate recordings.
• Knowledge and understanding of speech recognition techniques for paralinguistic elements (emotions, stress, lies).
• Knowledge and planning of the security management process for computer networks, personal computers and mobile terminals.
• Acquire advanced knowledge of traffic analysis and inspection of the operating history of equipment in a computer network.
• Understand the concepts of virtual network, authorization/authentication/registration, malware behaviour, etc.
• Skills in working with traffic inspection devices in computer networks.

Examples of dissertation topics from previous years:

• Protection of computer networks when handling data packets
• Forensic speaker recognition
• Dynamic signature authenticity recognition using neural networks
• Evolved voice communication interface under the Android operating system
• Machine learning of neural network hyperparameters
• Detection of non-verbal emotional prosodic expressions
• Method for controlling congestion and delays on the Internet
• Reverberation analysis in forensic audio recording forensics
• Traffic security systems for IoT devices
• User recognition using keystroke dynamics learning algorithms
• Separating speech segments from audio recordings using neural network models
• Trap system for analysing cyber attacks
• Applications of authentication and key distribution protocols in computer network security
• Automatic speech recognition system based on deep neural networks
• Learning algorithms for autonomous systems
• Method for automatic person recognition based on papillary fingerprints using machine learning techniques
• Identifying words of interest in speech transcripts and increasing the intelligibility of speech transcripts