Thesis type
Audio forensicsBestagini, Antonacci
Detection of text-to-speech algorithmsFull thesisNowadays, text-to-speech and voice conversion algorithms are able to produce very realistic speech signals, which can easily trick human ear. Moreover, this technology is in rapid evolution and it is not possible to take in account all new synthesis methods. It is necessary to develop effective synthetic speech detection systems able to work in open-set scenarios.
Musical acousticsAntonacci, Olivieri, PezzoliNearfield Acoustic Holography solver based on Physics-Informed Neural NetworkFull thesisPhysics-Informed Neural Networks (PINNs) are powerful optimizers that exploit the prior knowledge of physical laws as regularization agent also in unsupervised learning tasks.
This thesis aims at extending the recent approaches of data-driven Nearfield Acoustic Holography (NAH) with solutions based on PINN.
Musical acousticsAntonacci, Malvermi, PezzoliPhysics informed neural network for simulation of vibroacoustic phenomenaFull thesisRecently, deep learning strategies combined with a-priori knowledge given by partial differential equations proved to be effective in solving a wide range of physical problems. The goal of this thesis is to apply physics-informed paradigms to well-known vibroacoustic problems e.g., bending wave propagation, sound radiation etc. [Required knowledge: acoustics fundamentals, practical knowledge of neural networks]
Audio signal processing
Antonacci, PezzoliCharacterization and analysis of the directivity of sound sourcesFull thesis, Short thesisThe directivity is an inherent property of every sound source (e.g., a musical instrument). The goal of this thesis is to define suitable descriptors for the directivity of sound sources which can be used when comparing different directivities. Long or short thesis depends on the depth and novelty of the analysis. [Required knowledge: basic knowledge of statistical signal processing, spherical harmonics decomposition of sound field].
Audio signal processing
Antonacci, PezzoliDeep learning solution for localization of acoustic sources in the spherical harmonics domainFull thesisThe spherical harmonics representation of the sound field is a widely adopted description of spatial sound. The goal of this thesis is to devise deep learning solutions that exploit the spherical harmonics representation for the analysis of the acoustic field e.g., localization of acoustic sources. [Required knowledge: spherical harmonics decomposition of sound field, theoretical knowledge and pratical experience with deep learning]
Acoustic Scenes and EventsSarti, MezzaDetection of anomalous sounds in domestic environmentsFull thesisThe objective is the development of intelligent systems capable of monitoring a domestic environment to detect acoustic anomalies. The idea is to learn a suitable "normal state" for the monitored soundscape and to track the deviations thereof with minimal or no supervision.
Audio signal processing
Bernardini, GiampiccoloVirtual Analog Modeling, Audio Circuit Emulation, Physical Modeling Sound Synthesis through Wave Digital FiltersFull thesis
Audio signal processing
Bernardini, GiampiccoloImplementation of Wave Digital scattering matrices absorbing controlled sources Full thesisIn the Wave Digital domain, topological scattering junctions can absorb certain nonreciprocal elements. The thesis concerns the derivation of formulas for the efficient computation of their scattering matrices, such that they can be exploited for Virtual Analog applications.
Audio signal processing
Bernardini, GiampiccoloModeling of Nonlinear Piezoelectric LoudspeakersFull thesisDue to physical limitations, small-size transducers (e.g., piezoelectric and MEMS loudspeakers) are characterized by a poor acoustic response. The thesis concerns the derivation of nonlinear models of piezoelectric loudspeakers. Such models will be used in the future for developing algorithms for the transducer linearization, virtualization or compensation through circuital inversion.
Music informaticsSarti, Mezza, ComanducciAudio prediction for Networked Music Performance Full thesisIn recent years, live music performances over the Internet have become a reality. However, the inevitable network packets loss is still affecting the experience in a significant way, causing abrupt interruptions and glitches in the audio stream. In this respect, we aim at developing new deep learning-based techniques that would tackle this issue by predicting the next few milliseconds of audio, while ensuring the high quality and seamlessness of the reconstructed music signal.
Musical acousticsAntonacci, Pezzoli, MalvermiPrediction of new Frequency Response Functions through Convolutional Neural NetworksFull thesis, Short thesisRecently we have successfully applied CNNs to perform interpolation of Frequency Response Functions in rectangular plates also when dealing with random measurement locations. The thesis has as first target (short version) the extension of the method for real case scenarios (few initial measurements, violin plates, reconstruction of both real and imaginary part), then a new formulation exploting physics can be planned (long version).
Differential Microphone ArraysBernardini, AlbertiniEvaluation of phase, gain, and misplacement errors on arbitrary geometry differential microphone arraysFull thesisDifferential Microphone Arrays (DMAs) are small-sized microphone arrays known for their almost frequency-invariant beampatterns and superdirective responses. Recently developed arbitrary geometry DMAs represent a generalization of traditional DMAs, allowing to form spatial filters with any random microphone distribution. The aim of the Thesis is to model the effect of phase, gain and misplacement errors on the arbitrary geometry DMA performance, and develop a method to compensate such errors. Characterization of real sensors is a possibility.
Bernardini, AlbertiniAntiderivative Antialiasing for multi-port elements in the Wave Digital domainFull thesisAntiderivative Antialiasing is a recently proposed technique for mitigating aliasing distortion in virtual analog applications. The technique has attracted considerable attention due to its ability to reduce aliasing artifacts without resorting to high oversampling factors in the physical emulation of audio circuits. The approach has recently been extended to Wave Digital Filters with up to one-port nonlinear element. The goal of this thesis is to extend Antiderivative Antialiasing to multi-port nonlinear elements.
Differential Microphone ArraysBernardini, AlbertiniMultistage Beamforming with Differential Microphone ArraysFull thesisDifferential Microphone Arrays (DMAs) are small-sized microphone arrays known for their almost frequency-invariant beampatterns and superdirective responses. Recently, a two-stage spatial filtering (beamforming) using DMAs have been introduced. Multistage beamforming allows to obtain directivity patterns which are equal to the product of each spatial filtering stage beampattern. The aim of the Thesis is to develop a general N-stage beamformer and evaluate its properties in terms of directionality and robustness against noise.
Musical AcousticsGonzalez, AntonacciInfluence of thickness profile in archtop guitarsFull/short thesisThis thesis has two parts: modeling an archtop guitar parametrically in Fusion 360 and then realising simulations in COMSOL to study its acoustic response. Short and long thesis depend on the depth of the study.
Musical AcousticsAntonacci, Olivieri, PezzoliThe impact of reverberation for data-driven Nearfield Acoustic HolographyFull/short thesisRecent Deep Learning techniques proved the ability to infer the vibrational behavior in plates starting from acoustic measurements. This thesis aims at quantifying the performance of the current data-driven methods for Nearfield Acoustic Holography in the presence of reverberation (short thesis) and extending the neural network model to work in reverberant scenarios (long thesis).
Musical AcousticsAntonacci, Olivieri, PezzoliTransfer Learning techniques for Nearfield Acoustic Holography analysisFull/short thesisRecent data-driven based NAH methods can predict the vibrational behavior on sources from the acquisition of the radiated sound field. Nevertheless, these approaches are dependent on the training dataset used (i.e., acquisition setup and vibrational source). This thesis aims at extending the recent solutions with transfer learning strategies in order to tune the networks with different data and improving the model with specific physical priors to reconstruct the vibrational content with an unsupervised approach (long thesis). [knowledge of Deep Learning required]
Musical AcousticsGonzalez, AntonacciExperimental measurement and construction of violin top platesFull/short thesisThe aim of this thesis is twofold: measure the material properties of violin top plates and build violin top plates with certain material properties. For this the student will use a CNC router to build the plates and a experimental set up that measures the FRF of the plate to compute its material parameters. The goal is to be able to produce top plates with a defined mechanical response irrespective of the varying material parameters of the wood the top is made of. Experimental thesis in Cremona Campus, FEM modelling required, Fusion 360 optional.
Musical AcousticsGonzalez, AntonacciFEM simulation of cello bridgesShort thesisThe objective of this thesis is to study the structural and dynamical characteristics of different designs of Cello bridges. In particular, we will study the "classical" design and the "X" model created by Luthier L. Amorim. Knowledge of FEM is required, Fusion 360 would be a good extra.
Musical AcousticsGonzalez, AntonacciGuitar models catalog: design and simulationsLong thesisThe objective of this thesis is twofold: first to create a large catalog of guitar models in Fusion 360, including different bracing patterns and body shapes. After that is done simulations will be carried to understand how the different bracing patterns influence the vibrational and acoustic behaviour of the instrument.
Musical AcousticsGonzalez, AntonacciMultiphysics simulation of free reedsLong thesisThe bandoneon is fundamental part of the sound of Tango, yet there is very little scientific research on its sound. In this thesis we will study the multiphysics interaction of a free red by means of simulations, and hopefully compare the results with our experimental partners.
Musical AcousticsGonzalez, AntonacciRole of f-hole design in stress distribution on the violin top plateLong thesisThe role of the f-holes in violins is to let the sound vibrations leave the body of the instrument and reach the audience. However, cutting holes in the top plate weakens it. By cutting curves and circles, the instrument maker avoids creating the stress concentrations associated with sharp corners. The aim of this thesis is to study the stress behaviour in a violin soundboard for different f-hole designs. Comsol experience prefered.
Image and Video
Thesis type
Image/video forensicsBestagini, Mandelli, Cannas
Detect and localize image and video manipulations
FullImages and videos can be manipulated in many different ways (e.g., object insertion and removal, local retouching, laundering operations, etc.). We are interested in developing methods to detect and localize possible editing operations on images and videos.
Image/video forensicsBestagini, Mandelli, Cannas
Distinguish original videos from DeepFakesFullDeepFake videos can be maliciously spread online. We are interested in developing techniques to detect whether a video is a DeepFake or not, why a detector says a video is fake, and understand which DeepFake generation software has been used to create a video.
Image/video forensicsBestagini, Mandelli, Cannas
Assess the authenticity of satellite imagesFullSatellites can acquire visual data with different sensors. We are interested in developing techniques that verify whether an overhead image has been edited or not.
Image/video forensicsBestagini, Mandelli, Cannas
Forensic analysis of scientific imagesFullScientific publications in the life science area typically contain charcateristic kinds of images to showcase the achieved results (e.g., western blots, microscopy acquisitions, etc.). As these images differ from natural photographs, we are interested in developing novel techniques to detect possible scientific image forgery operations.
Image/video processingMarconAnalysis of hyperspectral X-ray images for food quality assessmentFullX-ray acquisitions are beneficial in food contaminant analysis as they can detect both metallic and non-metallic objects. The goal of a thesis in this field is to develop processing techniques that help dealing with X-ray acquisition in food contaminant detection applications (e.g., denoising, detection, segmentation, etc.).
Thesis type
Tubaro, LipariImproving Full Waveform Inversion with CNNsFull/shortFull Waveform Inversion reconstructs the subsurface velocities from a set of measurements. It is very expensive, time-consuming and prone to a number of tips and tricks for avoiding local minima, numerical instability and optimization errors.
Tubaro, LipariDenoising and Interpolation of seismic data through CNNsFull/shortThe amount of data is constantly increasing and the areas of interest are more and more complex to analyze. Moreover, they require a subsurface mapping at increasingly higher resolution and higher fidelity. Can CNNs help this process?
Tubaro, LipariMachine Learning guided Seismic InterpretationFull/shortHuman experts visually inspects seismic images looking for subsurface features. On the other hand, Machine Learning techniques have proven to be effective in image segmentation (i.e., recognizing objects and targets from a set of pixels). Can we merge these two worlds?
Tubaro, LipariPhysics-aware Transfer LearningFull/shortCan a neural network learn Physics?
Tubaro, LipariRegularizing Traveltime Tomography via Machine LearningFull/shortTraveltime Tomography is an important tool for localizing subsurface events, both in medical and geophysical imaging. The tomography problem is a ill-posed and ill-conditioned inverse problem, which suffers a lot from numberical instability, local minima, and noise. Here's where the machine learning comes into play!

Currently on-going

Expand list
Image forensicsBestagini, MandelliManipulation detection for scientific imagesGiovanni Zanocco
Video processingBestagini, RedondiAutomatic video analysis of badminton matchesIvan Motasov
Audio signal processingBernardini, GiampiccoloVirtual Bass Enhancement for Small-Size TransducersFabio Spreafico
Space-time audioBernardini, GiampiccoloSynthesis of Room Acoustics via Scattering Delay Networks using Automatic DifferentiationFrancesco Boarino
Audio signal processingBernardini, GiampiccoloModeling nonlinear two-ports by means of Vector Waves in the Wave Digital DomainDavide Marin Pasin
Musical AcousticsAntonacci, OlivieriTowards white-box data-driven methods for Near-field Acoustic HolographyHagar Kafri
Video forensicsBestagini, CannasDeepfake video detection through multi-look analysisAdriano Bonfantini
Audio signal processingBernardini, GiampiccoloWave Digital Newton-Raphson Method for Virtual Analog ModelingLuca Gobbato
Space-time audioAntonacci, PezzoliSound source localization in the spherical harmonicsLuca Cattaneo
Music informaticsSarti, Mezza, BernardiniUnsupervised selection of harmonic complexity metricsGiorgio De Luca
Music informaticsZanoni, BorrelliSocial interaction based music recommendation systemCarlo Pulvirenti
Music informaticsSarti, BorrelliConnecting NN to bio-metric signalsJoep Rene Wulms
Musical AcousticsRipamonti, Malvermi, GonzalezExperimental Validation for data-driven Near-field Acoustic HolographyAlessio Lampis
Musical AcousticsAntonacci, MalvermiImproved sensors for low-cost Vibrometric KitFabio Guarnieri
Musical AcousticsSarti, Paoletti, Adali, MalvermiAcoustic Characterization of materialsMarco Donzelli
Musical AcousticsGonzalez, AntonacciRandom variation of guitar bracingsMattia Vanessa
Musical acousticsGonzalez, AntonacciMetamaterials for guitarmakingGabriele Marelli, Mattia Lercari
Musical acoustics / AIGonzalez, AntonacciAI-powered pick up: making guitars sound great againEmanuele Voltini
Music Informatics Sarti, Turchet (UNITN), ComanducciAdaptive Metronomes using Haptic Devices for Networked Music PerformancesPaolo De Santis
Space-time audioAntonacci, ComanducciPersonalized Sound Zone Generation using Deep LearningRoberto Alessandri

Past (from 2017)

Expand list
Music Informatics Zanoni, ComanducciDeep Learning-based Timbre TransferSilvio Pol
Audio signal processingAntonacciA perceptual evaluation of sound field reconstruction algorithmsMiriam Papagno
Audio signal processing
Bernardini, GiampiccoloCharacterization of Small-Size Loudspeakers for Mobile ApplicationsSamuele Buonassisi
Image forensicsBestagini, CannasEnhanced Amplitude SAR Imagery Splicing Localization through Land Cover Mapping TechniquesEmanuele Intagliata
GeophysicsBestagini, LipariSalt Segmentation of Geophysical Images through Explainable CNNsFrancesco Maffezzoli
Audio forensicsBestagini, BorrelliA metric learning approach for splicing localization based on synthetic speech detectionFrancesco Castelli
Audio forensicsBestagini, BorrelliCombining automatic speaker verification and prosody analysis for synthetic speech detectionLuigi Attorresi
Music informaticsBestagini, CuccovilloSpeech fingerprinting and matching for content retrievalLaura Colzani
Video forensicsBestaginiA CNN-based detector for video frame-rate interpolationSimone Mariani
Image/video processingBestaginiAudio-video techniques for the analysis of players behaviour in Badminton matchesSamuele Bosi
Video forensicsBestagini, MandelliForensic detection of deepfakes generated through video-to-video translationCarmelo Fascella
Audio signal processing
Bernardini, Mezza, GiampiccoloWave Digital Filter Modeling of Audio Circuits with Hysteresis Nonlinearities using Neural NetworksOliviero Massi
Music informaticsAntonacci, Pezzoli, ComanducciDeep Prior Audio InpaintingFederico Miotello
Audio signal processingBestagini, BuccoliLow-latency speaker recognitionFrancesco Salani
Video forensicsBestagini, BonettiniA Data Driven Approach to Deepfake Detection via Feature Analysis Based on Limited Data Bingyang Hu
Space-time audioAntonacci, Borrelli, BorraBeamforming and Speaker Identification through Deep Neural Networks Matteo Scerbo
Music informaticsSarti, BorrelliHarmonic complexity estimation of jazz musicGiovanni Agosti
Audio forensicsAntonacci, BorrelliA model selection method for room shape classification based on mono speech signalsGabriele Antonacci
Audio forensicsBestaginiAudio splicing detection and localization based on recording device cuesDaniele Ugo Leonzio
Audio forensicsBestaginiSpeaker-Independent Microphone Identification via Blind Channel Estimation in Noisy ConditionAntonio Giganti
Audio forensicsBestagini, BorrelliSynthetic Speech Detection through Convolutional Neural Networks in Noisy EnvironmentsEleonora Landini
Audio forensicsBestagini, Borrelli, SalviSynthetic speech detection based on sentiment analysisEmanuele Conti
Multimedia forensicsBestagini, Salvi, BorrelliAudio-video deepfake detection through emotion recognitionJacopo Gino
Audio signal processingSarti, Giampiccolo, BernardiniParallel Wave Digital Implementations of Nonlinear Audio CircuitsNatoli Antonino
Musical AcousticsAntonacci, MalvermiData driven methods for frequency response functions interpolationMatteo Acerbi
Audio forensicsBestagini, MandelliTime-Scaling Detection in Audio RecordingsMichele Pilia
Audio forensicsBestagini, BorrelliSpeech Intelligibility Parameters Estimation Through Convolutional Neural NetworksMattia Papa
Audio forensicsAntonacciClosed and open set classification of real and AI synthesised speechMichelangelo Medori
Audio forensicsAntonacciAn approach to room volume estimation from single-channel speech signals based on neural networksCastelnuovo Carlo
Audio forensicsBestaginiAudio Splicing Detection and Localization Based on Acoustic CuesCapoferri Davide
Audio processingSarti, ComanducciAudio frame reconstruction from incomplete observations using Deep Learning techniquesSchils Minh Cédric
Audio processingSarti, BernardiniWave Digital Modeling and Simulation of Nonlinear Electromagnetic CircuitsGiampiccolo Riccardo
Audio processingSarti, BernardiniAntiderivative Antialiasing in Nonlinear Wave Digital FiltersAlbertini Davide
Audio processingSarti, BernardiniWave Digital Implementation of Nonlinear Audio Circuits based on the Scattering Iterative MethodProverbio Alessandro
Audio processingAntonacciA system for super resolution vibrometric analysis through convolutional neural networksCampagnoli Chiara
Audio processingAntonacciDevelopment of a low-cost platform for acoustic and vibrometric analysis on lutherie products, with a special focus on the estimation of the elastic parameters of the tonewoodVilla Luca
Audio processingBestaginiDNN based post-filtering for quality improvement of AMR-WB decoded speechGupta Kishan
Audio processingSartiStudio sull'implementazione degli algoritmi per il musical instruments ed il sound reinforcement basato su un processore multicoreAretino Michele
Audio processingSarti, BernardiniModeling nonlinear 3-terminal devices in the wave digital domainVergani Alessio Emanuele
ForensicsBestaginiConvolutional and recurrent neural networks for video tampering detection and localizationCannas Edoardo Daniele
ForensicsBestaginiA study on Bagging-Voronoi algorithm for tampering localizationCereghetti Corinne Elena
ForensicsBestaginiJPEG-based forensics through convolutional neural networksBonettini Nicolò
ForensicsBestaginiAnalysis of different footprints for JPEG compression detectionChen Ke
GeophysicsBestaginiLandmine detection on GPR data employing convolutional autoencoderTesta Giuseppe
Image and videoMarcon, ParacchiniA novel tomographic approach for an early detection of multiple myeloma progressionAndrea Leggio
Image and videoMarcon, ParacchiniLimited angle computed tomography reconstruction with deep learning enhancementErbol Kasenov, Girolamo Gerace
Image and videoMarconUpper body postural assessment for common dentistry visual aidsTrotta Emilio
Image and videoTubaroReal-time tracking of electrode during deep-brain surgeryDilauro Valerio
Image and videoMarconAnalytical estimation of the error on the radius of industrial pipesLazzarin Sara
Image and videoMarcon3D reconstruction from stereo video acquired from odontoiatric microscopeSpatafora Leonardo
Image and videoMarconDenoising and classification of hyperspectral X-ray images for food quality assessmentRe Marco
Image and videoMarconA computer vision approach for assessment of dental bracket removalBehnami Arezoo
Image and videoMarconSistema per il rilevamento automatico di contaminanti alimentari basato su immagini iperspettraliRamoni Francesco
Image and videoMarconPostural assessment in dentistry by computer visionPignatelli Nicola
Multimedia forensicsBestagini, MandelliA Multi-Modal Approach to Forensic Audio-Visual Device IdentificationDavide Dal Cortivo
Music informaticsSarti, Bernardini, Borrelli, MezzaEstimating Harmonic Complexity of Chord Sequences using Transformer NetworksCecilia Morato
Music informaticsZanoni, ComanducciModeling Harmonic Complexity in Automatic Music Generation using Conditional Variational AutoencodersDavide Gioiosa
Music informaticsSarti, Borrelli, ComanducciCellular music : a novel music-generation platform based on an evolutionary paradigmMatteo Manzolini
Music informaticsSarti, BorrelliMusic emotion detection. A framework based on electrodermal activities.Gioele Pozzi
Music informaticsSarti, ComanducciTechniques for mitigating the impact of latency in
Networked Music Performance (NMP) through adaptive metronomes
Battello Riccardo
Music information retrievalSartiMusical instrument recognition: a transfer learning approachMolgora Andrea
Music information retrievalSartiUnsupervised domain adaptation for deep learning based acoustic scene classificationMezza Alessandro Ilic
Music information retrievalAntonacciAn investigation of piano transcription algorithm for jazz musicMarzorati Giorgio
Music information retrievalSartiAutomatic playlist generation using recurrent neural networkIrene Rosilde Tatiana
Music information retrievalSartiA personalized metric for music similarity using Siamese deep neural networksSala Federico
Music information retrievalSartiLearning a personalized similarity metric for musical contentCarloni Luca
Music information retrievalSartiBeat tracking using recurrent neural network : a transfer learning approachFiocchi Davide
Music information retrievalSartiPython-based framework for managing a base of complex data for music information retrievalAvocone Giuseppe
Music information retrievalSartiIndividual semantic modeling for music information retrievalAnsidei Pietro
Music information retrievalSartiChord sequences : evaluating the effect of complexity on preferenceFoscarin Francesco
Music information retrievalSartiAudio features compensation based on coding bitrateTavella Maria Stella
Musical AcousticsAntonacciModal analysis and optimization of the top plate of string instruments through a parametric control of their shapeSalvi Davide
Musical AcousticsAntonacci, Pezzoli, Malvermi An approach for Near-field Acoustic Holography based on Convolutional AutoencodersOlivieri Marco
Space-time audioAntonacci, BorraA parametric approach to virtual miking with distributed microphone arraysMarco Langè
Space-time audioAntonacci, Pezzoli, Borra, BernardiniA Deep Prior Approach to Room Impulse Response InterpolationDavide Perini
Space-time audioAntonacci, ComanducciInterpreting Deep Neural Networks Models for Acoustic Source Localization using Layer-wise Relevance PropagationAlessandro Montali
Space-time audioAntonacci, Borra, BernardiniAnalysis of Uniform Linear Arrays of Differential MicrophonesBertuletti Ivan
Space-time audioSartiA geometrical method of 3D sound spatialization for virtual reality applicationsIamele Jacopo
Space-time audioAntonacciConvolutional neural networks applied to space-time audio processing applicationsComanducci Luca
Space-time audioCancliniDenoising in the spherical harmonic domain of sound scenes acquired by compact arraysBorrelli Clara
Space-time audioAntonacciSimulazione di sistemi complessi. Case study : l'altoparlante a trombaMoscara Francesco
Space-time audioSarti, BernardiniSteerable differential microphone arraysLovatello Jacopo
Space-time audioAntonacciA plenacoustic approach to sound scene manipulationPicetti Francesco
Space-time audioAntonacciReconstruction of the soundfield in arbitrary locations using the distributed ray space transformPezzoli Mirco
Space-time audioSartiA method for HRTF personalization : weighted sparse representation synthesis of HRTFsZhu Mo
Space-time audioAntonacciRobust parametric spatial audio processing using beamforming techniquesMilano Guendalina
Space-time audioAntonacciEstimation of singing voice quality through microphone in air and contact microphoneLandini Roberta
Musical AcousticsAntonacci, MalvermiMechanical parameter estimation for vibrometric analysis and development of a low-cost platform for violin makingFederico Simeon
Space-time audioAntonacci, Comanducci3D audio with irregular microphone setups using deep learningDavide Mori