Biometrics and Multimedia Forensics

The educational objectives of the Biometrics and Multimedia Forensics teaching are as follows:

1. Knowledge and understanding: the student will learn the knowledge of biometrics and digital forensics techniques, with particular attention to forensic phonetics and mobile forensics techniques.
2. Ability to apply knowledge and understanding: the knowledge of an applicative nature, acquired through the exercises and the laboratory, will concern the ability to use the main tools adopted in digital forensics; the student will acquire problem solving and teamwork skills.
3. Making judgments: teaching will stimulate autonomy of judgment and assessment of the conditions in which to apply digital forensics techniques and tools.
4. Communication skills: teaching is based on the use of a language specific to digital forensics studies which will become the basis of the communication activity set up by the student;
5. Learning skills: learning skills will be stimulated by critical knowledge of the topics covered in teaching made possible by attendance at lectures, laboratory activities and the study of reference texts.

The course includes a combination of lectures and laboratory sessions for a total of 58 hours and is jointly taught by Prof. Francesco Beritelli and Prof. Roberta Avanzato.

Signal Theory, Fundamentals of Telecommunications or Digital Communications.

Attendance at the lessons is not compulsory, however, it should be noted that only by participating in the laboratory activities will it be possible to acquire up to 3 points of the final grade, keeping in mind that the oral evaluation is up to 27 points (see "Methods of verification of learning").

1. Introduction to the course (2 hours): Topics in digital forensics. Differences between multimedia forensics and computer forensics. Legal aspects. Technical consultancy and expert witness reports in criminal proceedings. The roles of the court-appointed expert (CTU) and the party-appointed expert (CTP).

2. Biometric identification techniques (6 hours): Introduction to biometric systems. Authentication. Morphological and behavioral biometric identifiers. Architecture of a biometric system. Statistical methods. Matching and decision techniques. Verification and evaluation of system performance, usability, and scalability. Performance metrics (FAR, FRR, ROC, EER, DET). Multimodal biometric systems.

3. Audio forensics (6 hours): Main techniques for audio signal processing and compression. Speech coding standards. Forensic phonetics. Audio wiretapping. Audio signal enhancement techniques. Voice identification techniques: auditory methods, semi-automatic methods (formants, LPC), automatic methods (MFCC, GMM, i-vectors). Voice imitation (deepfake). Anti-spoofing methods and voice liveness detection. Audio signal transcription.

4. Video forensics (6 hours): Image and video signal processing. Main compression techniques and coding standards. Video forensic techniques. Generation techniques: GANs, autoencoders, diffusion models. Forensic indicators of deepfakes (facial inconsistencies, blinking, eye gaze, texture). Detection of manipulations, alterations, and deepfakes. AI and Computer Vision for security in Smart Cities.

5. Mobile forensics (3 hours): Introduction to mobile forensics. Principles and procedures. Preservation. Acquisition. Inspection and analysis. Reporting. Main tools.

6. Geolocation in mobile radio systems (5 hours): Geolocation services and systems. Localization techniques. Triangulation and trilateration using Angle of Arrival (AoA), Time of Arrival (ToA), and Time Difference of Arrival (TDoA) algorithms, hybrid techniques. Methods for reducing localization errors.

7. Laboratory (30 hours): Main tools for forensic phonetics: PRAAT, Idem, Nuance Forensic. Use of Python/MATLAB libraries for facial and voice recognition. Projects on real datasets (e.g., FERET, LFW, VoxCeleb, FVC, PhysioNet).

1. Teacher's notes
2. "Introduction to Biometrics“, Anil K. Jain , Arun A. Ross , Karthik Nandakumar , Thomas Swearingen, Springer, 2nd Ed., 2025
3. "Multimedia Forensics", Husrev Taha Sencar, Luisa Verdoliva, Nasir Memon, Springer, 2022
4. “Guidelines on Mobile Device Forensics", Rick Ayers, Sam Brothers, Wayne Jansen, NIST, Special Publication 800-10 Revision 1, May 2014

The assessment concerns an optional term paper or project and 2–3 oral questions, for a maximum of 27 points. Based on attendance in the laboratory sessions, up to 3 additional points may be added to the final grade. Laboratory sessions will generally range from a minimum of 4 to a maximum of 8.

The questions cover all topics of the course.