Wednesday, 17/9

Keynote 1 - Synthesia Avatars and Vision + Language Research

Invited speaker: Vittorio Ferrari

Abstract: The first part of the talk will be about Synthesia, which enable users to create a video simply by selecting an avatar and entering the words it should say. Synthesia then generates a full performance, including speech, animating the avatar to match that speech, and rendering into a video. In the talk I will chart the technical evolution of Synthesia's avatars towards ever more expressive, engaging, and photorealistic avatars. The second part of the talk will be about our research in the "Vision and Language" area. I will cover recent works on open vocabulary segmentation with foundation models, multimodal image and video annotation with voice&mouse interfaces, and encyclopedic VQA.

Oral session 1.A - Computer Vision

11:30 - 11:50

FSMDet: Vision-guided feature diffusion for fully sparse 3D detector

Tianran Liu (University of Ottawa)*; Morteza Pasandi (University of Ottawa); Robert Laganiere (University of Ottawa)

11:50 - 12:10

Leveraging Multi-View Weak Supervision for Occlusion-Aware Multi-Human Parsing

Laura Bragagnolo (University of Padova)*; Matteo Terreran (University of Padova); Leonardo Barcellona (University of Padova); Stefano Ghidoni (University of Padova)

12:10 - 12:30

Fault tolerant multi-object tracking via temporal consistency filtering

Rosario Milazzo (Politecnico di Torino)*; Sophie Fosson (Politecnico di Torino); Lia Morra (Politecnico di Torino); Luca Sterpone (Politecnico di Torino)

12:30 - 12:50 

Zero-Shot Neural Architecture Search for Efficient Deep Stereo Matching

Alessio Mingozzi (Alma Mater Studiorum - University of Bologna)*; Stefano Mattoccia (Alma Mater Studiorum - University of Bologna); Matteo Poggi (Alma Mater Studiorum - University of Bologna); Fatma Güney (KUIS AI Center, Koç University)

Oral session 1.B - Learning for Cultural Heritage

14:30 - 14:50

Successful 3D Reconstructions of Underwater World War II Wreckage

Jaren Gerdes (Alutiiq, LLC)*; Jane Berk (Alutiiq, LLC); Shane Wechsler (San Diego State University); Mark Bilinski (NIWC Pacific); Mike Hess (Alutiiq, LLC); Ray Provost (NIWC Pacific); Angaar Wasimi (NIWC Pacific)

14:50 - 15:10

Beyond Linear Bottlenecks: Spline-Based Knowledge Distillation for Culturally Diverse Art Style Classification

Abdellah Sellam (University of Salento)*; Salah Eddine bekhouche (University of the Basque Country); Cosimo Distante (CNR-ISASI); Abdelmalik Taleb-Ahmed (Université Polytechnique Hauts-de-France)

15:10 - 15:30

Benchmarking Content-Based Puzzle Solvers on Corrupted Jigsaw Puzzles

Richard Dirauf (Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU))*; Florian Wolz (Siemens Healthineers AG); Dario Zanca (Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU)); Björn Eskofier (Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU))

Oral session 1.C - Video Analysis

17:00 - 17:20

VidCLearn: A Continual Learning Approach for Text-to-Video Generation

Luca Zanchetta (Sapienza)*; Lorenzo Papa (Sapienza); Luca Maiano (Sapienza); Irene Amerini (Sapienza)

17:20 - 17:40

Mamba-OTR: a Mamba-based Solution for Online Take and Release Detection from Untrimmed Egocentric Video

Alessandro Sebastiano Catinello (University of Catania); Giovanni Maria Farinella (University of Catania); Antonino Furnari (University of Catania)*

17:40 - 18:00

Ego-Exo Object Correspondence by SAM2 and Cross-View Prompting

Devis Salierno (University of Udine); Matteo Dunnhofer (University of Udine)*; Christian Micheloni (University of Udine)

Poster session 1

SegDT: A Diffusion Transformer-Based Segmentation Model for Medical Imaging

Salah Eddine Bekhouche (University of Basque Country)*; Gaby Maroun (University of Basque Country); Fadi Dornaika (University of Basque Country); Abdenour Hadid (Sorbonne University Abu Dhabi)

Unsupervised segmentation for reliable estimation of stratum corneum thickness using LC-OCT

Raoul Missodey (Laboratoire PRISME - Université d'Orléans)*; Remy Leconge (Laboratoire PRISME); Samuel Ralambondrainy (Lvmh Recherche); Sylvie  Treuillet (Laboratoire PRISME); Yves Lucas ( Laboratoire PRISME); Jean-Hubert Cauchard (Lvmh Recherche); Franck  Bonnier (Lvmh Recherche)

Decoding Human Emotions: A Robust Ensemble Approach to Facial Emotion Recognition

Alexandru Berindeie (West University of Timisoara)*

Deep Multi-Band EEG Learning for Motor Imagery Classification with Dry Electrodes

Hamza Amrani (University of Milano-Bicocca); Daniela Micucci (University of Milano-Bicocca); Paolo Napoletano (University of Milano - Bicocca)*

SVGauge: Towards Human-Aligned Evaluation for SVG Generation

Leonardo Zini (University of Modena and Reggio Emilia); Elia Frigieri (University of Modena and Reggio Emilia); Sebastiano Aloscari (University of Modena and Reggio Emilia); Marcello Generali (Doxee S.p.A.); Lorenzo Dodi (Doxee S.p.A.); Robert Dosen (Doxee S.p.A.); Lorenzo Baraldi (University of Modena and Reggio Emilia)*

Revisiting dictionaries of key poses for action representation

Matteo Moro (University of Genoa and MaLGa Center)*; Federico Figari Tomenotti (University of Genoa and MaLGa Center); Nicoletta Noceti (University of Genoa and MaLGa Center); Francesca Odone (University of Genoa and MaLGa Center)

DeepCS-TRD, a Deep Learning-based Cross-Section Tree Ring Detector

Henry Marichal (Facultad de Ingenieria, Universidad de la Republica)*; Victoria Casaravilla (Sede Rivera, CENUR Noreste, Universidad de la República); Candice Power (Ecoinformatics and Biodiversity, Department of Biology, Aarhus University); Karolain Mello (Sede Rivera, CENUR Noreste, Universidad de la República); Joaquin Mazarino (Sede Paysandú, CENUR Litoral Norte, Universidad de la República); Christine Lucas (Sede Paysandú, CENUR Litoral Norte, Universidad de la República); Ludmila Profumo (Sede Rivera, CENUR Noreste, Universidad de la República); Diego Passarella (Centro Universitario Regional Noreste, Universidad de la República, Tacuarembo); Gregory Randall (Facultad de Ingenieria, Universidad de la Republica)

Unveiling Truth Through Gestures: A Multimodal Deepfake Detection System

Vishruth Rao (Eastlake High School)*

Comparative Analysis of Lightweight Vision Models on Embedded Accelerator Devices

Greta Corti (Politecnico di Milano)*; Christian Veronesi (Politecnico di Milano); Pietro Bartoli (Politecnico di Milano); Sara Pido (Politecnico di Milano); Andrea Giudici (Politecnico di Milano); Francesca Palermo (EssilorLuxottica); Diana Trojaniello (EssilorLuxottica); Matteo Matteucci (Politecnico di Milano); Simone Mentasti (Politecnico di Milano)

Ensemble-Enhanced Semantic Segmentation for Efficient Waste Recycling

Mohammadhossein Akbari Moafi (University of Padova)*; Horst Possegger (Graz University of Technology)

Accelerating YOLO with EOBranch: An Early Exit Approach for Adaptive Object Detection

Dick Scholte (Eindhoven University of Technology)*; Matthijs Zwemer (ViNotion B.V.); Egor Bondarev (Eindhoven University of Technology)

MATE: Multimodal Agents that Talk and Empathesize

Niyati Rawal (University of Modena and Reggio Emilia)*; Matteo Xia (Omitech S.R.L.); David Tessaro (Omitech S.R.L.); Lorenzo Baraldi (University of Modena and Reggio Emilia); Rita Cucchiara (University of Modena and Reggio Emilia)

Multiple Sclerosis Classification via Random Forest Distances Robust to Missing Data

Antonella Mensi (University of Verona)*; Vincenzo Di Maria (University of Verona); Elena Barusolo (University of Verona); Roberta Magliozzi (Imperial College London); Manuele Bicego (University of Verona)

Efficient oblique stripe noise detection and removal using Hough transform and guided filter

Ayoub Boutemedjet (Ecole Militaire Polytechnique)*

Predicting Soccer Penalty Kick Direction Using Human Action Recognition

David Freire-Obregón (Universidad de Las Palmas de Gran Canaria)*; Oliverio Santana (Universidad de Las Palmas de Gran Canaria); Javier Lorenzo-Navarro (Universidad de Las Palmas de Gran Canaria); Daniel Hernández-Sosa (Universidad de Las Palmas de Gran Canaria); Modesto Castrillón-Santana (Universidad de Las Palmas de Gran Canaria)

Hierarchical Vision-Language Retrieval of Educational Metaverse Content in Agriculture

Ali Abdari (University of Udine, University of Naples Federico II)*; Alex Falcon (University of Udine); Giuseppe Serra (University of Udine)

Depth and Event-Based Approaches for Human Detection and Pose Estimation

Hao Quan (Politecnico di Milano)*; Milad Goudarzi (Politecnico di Milano); Chiara Plizzari (Politecnico di Milano); Francesca Palermo (EssilorLuxottica); Simone Mentasti (Politecnico di Milano); Diana Trojaniello (EssilorLuxottica); Matteo Matteucci (Politecnico di Milano)

Prompt Based Video Colorization

Silvia Dani (University of Florence); Tiberio Uricchio (University of Pisa); Lorenzo Seidenari (University of Florence)*

D2R2: Deep-Learning based Real-time Dense object Reconstruction 

Marco Lombardi (University of Brescia); Mattia Savardi (University of Brescia)*; Alberto Signoroni (University of Brescia)

FS-SAM2: Adapting Segment Anything Model 2 for Few-Shot Semantic Segmentation via Low-Rank Adaptation

Bernardo Forni (University of Pavia)*; Gabriele Lombardi (ARGO Vision); Federico Pozzi (ARGO Vision); Mirco Planamente (ARGO Vision)

Deep Multimodal Fusion for 2D-3D Vineyard Scene Understanding

Mirko Usuelli (Politecnico di Milano); Cristian Sbrolli (Politecnico di Milano)*; Alessio Braccini (Politecnico di Milano); Matteo Frosi (Politecnico di Milano); Matteo Matteucci (Politecnico di Milano)

Disentangled Latent Spaces Facilitate Data-Driven Auxiliary Learning

Geri Skenderi (Bocconi University)*; Luigi Capogrosso (University of Verona); Andrea Toaiari (University of Verona); Matteo Denitto (HUMATICS - SYS-DAT Group); Franco Fummi (University of Verona); Simone Melzi (University of Milano-Bicocca)

VS-Sim: A Synthetic Dataset for Viewpoint Shift Robustness

Riccardo Turra (Verizon Connect); Matteo Simoncini (Verizon Connect); Henrique Piñeiro Monteagudo (Università di Bologna; Verizon Connect)*; Aurel Pjetri (Verizon Connect; University of Florence); Samuele Salti (Università di Bologna); Leonardo Taccari (Verizon Connect)

Deep Transductive Learning for Person Re-Identification

Hamzah Alghamdi (University of Ottawa)*;  Wassim  El Ahmar (University of Ottawa); Robert Laganière (University of Ottawa)

An Investigation of Ear-EEG Signals for a Novel Biometric Authentication System

Danilo Avola (Sapienza University of Rome); Giancarlo Crocetti (St. Johns University); Gian Luca Foresti (University of Udine); Daniele Pannone (Sapienza University of Rome)*; Claudio Piciarelli (University of Udine); Amedeo Ranaldi (Sapienza University of Rome); Davide Vittucci (Sapienza University of Rome)

Inpainting of Ancient Chinese Character Rubbings via Generative Adversarial Network

Xiangheng Wang (Ritsumeikan University)*

Thursday, 18/9

Keynote 2 - Scene Understanding from Limited Supervision

Invited speaker: Stefan Roth

Abstract: Scene understanding has made huge strides over the past decade though much of it required the use of large human-annotated datasets. I will discuss how we can understand 2D and even 3D scenes with little to (virtually) no annotation effort, leveraging various forms of consistency. I will show applications to semantic segmentation, panoptic segmentation, and 3D semantic scene completion. I will also discuss how we can learn latent compositional representations of visual scenes that allow for a multitude of downstream tasks, including object discovery, compositional generation, and more.

Industrial panel

The panel will host leading figures from the academia and industry to discuss the state-of-the-art in AI applied to computer vision.

Oral session 2.A - Biomedical, Assistive Technology, Social Signal Processing

10:30 - 10:50

MADPOT: Medical Anomaly Detection with CLIP Adaptation and Partial Optimal Transport

Mahshid Shiri (University of Verona); Cigdem Beyan (University of Verona)*; Vittorio Murino (University of Verona)

10:50 - 11:10

Enhancing Testicular Ultrasound Image Classification through Synthetic Data and Pretraining Strategies

Nicola Morelli (Università degli Studi di Modena e Reggio Emilia); Kevin Marchesini (Università degli Studi di Modena e Reggio Emilia); Luca Lumetti (Università degli Studi di Modena e Reggio Emilia); Daniele Santi (Università degli Studi di Modena e Reggio Emilia); Costantino Grana (Università degli Studi di Modena e Reggio Emilia); Federico Bolelli (Università degli Studi di Modena e Reggio Emilia)*

11:10 - 11:30

A Semantically-Aware Relevance Measure for Content-Based Medical Image Retrieval Evaluation

Xiaoyang Wei (Université Paris Cité)*; Camille Kurtz (Université Paris Cité); Florence Cloppet (Université Paris Cité)

11:30 - 11:50

AGCD-Net: Attention Guided Context Debiasing Network for Emotion Recognition

Varsha Devi (CESI LINEACT); Amine Bohi (CESI LINEACT)*; Pardeep Kumar (INRIA, Sophia Antipolis)

11:50 - 12:10

Efficient Calisthenics Skills Classification through Foreground Instance Selection and Depth Estimation

Antonio Finocchiaro (University of Catania)*; Giovanni Maria Farinella (University of Catania); Antonino Furnari (University of Catania)

Oral session 2.B - Learning with Large Models

14:00 - 14:20

How Far Can Off-the-Shelf Multimodal Large Language Models Go in Online Episodic Memory Question Answering?

Giuseppe Lando (University of Catania); Rosario Forte (University of Catania); Giovanni Maria Farinella (University of Catania); Antonino Furnari (University of Catania)*

14:20 - 14:40

7Bench: a Comprehensive Benchmark for Layout-guided Text-to-image Models

Elena Izzo (University of Padova)*; Luca Parolari (University of Padova); Davide Vezzaro (University of Padova); Lamberto Ballan (University of Padova)

14:40 - 15:00

Automatic benchmarking of Large Multimodal Models via iterative experiment programming

Alessandro Conti (University of Trento)*; Enrico Fini (Independent Researcher); Paolo Rota (University of Trento); Yiming Wang (FBK); Massimiliano Mancini (University of Trento); Elisa Ricci (University of Trento)

15:00 - 15:20

Unsupervised Anomaly Localization In the Wild via Token Optimization and Test-Time Score Correction

Naoki Kawamura (Toshiba Corporation)*; Gaku Minamoto (Toshiba Corporation); Tomohiro Nakai (Toshiba Corporation); Satoshi Ito (Toshiba Corporation); Osamu Yamaguchi (Toshiba Corporation); Takahiro Takimoto (Toshiba Corporation)

Poster session 2

Explainable Machine Learning for Earthquakes: SHAP Interpretation of CNNs to Distinguish Seismic Spectrograms of Foreshocks and Aftershocks

Michele Magrini (Sapienza University of Rome)*; Francesco Marrocco (Sapienza University of Rome); Laura Laurenti (ETH Zurich); Gabriele Paoletti (Sapienza University of Rome); Elisa Tinti (Sapienza University of Rome); Chris Marone (Sapienza University of Rome)

Physics-Based 3D Simulation for Synthetic Data Generation and Failure Analysis in Packaging Stability Assessment

Samuel Seligardi (University of Parma); Pietro Musoni (University of Parma)*; Eleonora Iotti (University of Parma); Gianluca Contesso (ACMI S.p.A.); Alessandro Dal Palù (University of Parma)

Ego and exo views for an object-level human behavior analysis and understanding through tracking in retail spaces

Alessandro Sebastiano Catinello (University of Catania); Matteo Dunnhofer (University of Udine); Giovanni Maria Farinella (University of Catania); Emanuele Frontoni (University of Macerata); Antonino Furnari (University of Catania); Christian Micheloni (University of Udine); Marina Paolanti (University of Macerata); Rocco  Pietrini (Università Politecnica delle Marche)*; Devis Salierno (University of Udine); Lorenzo Stacchio (University of Macerata); Asfand Yaar (University of Catania)

A SAM-based Automated Schistocyte Detection Pipeline in Peripheral Blood Smear Images

Ahmed Bensaid (University of Cagliari); Andrea Loddo (University of Cagliari)*; Lorenzo Putzu (University of Cagliari); Cecilia Di Ruberto (University of Cagliari)

A Study on Multimodal Foundation Models for Affective Video Prediction

Francesco Agnelli (University of Milan)*; Mirko Ditroia (University of Milan); Giorgio Blandano (University of Milan); Omar Ghezzi (University of Milan); Alessandro D'amelio (University of Milan); Raffaella Lanzarotti (University of Milan)

Business Activity Classification Extraction from Commercial Footage: A Multimodal LLM Approach Based on CNAE (Comparable to NACE and NAICS)

Rogerio Lopes (Itaú Unibanco)*; Dibio Borges (Universidade de Brasília)

Multi-Frame Alignment for Video Super-Resolution using Attention

Marco Di Rienzo (University of Florence); Matteo Bruni (Small Pixels); Leonardo Galteri (Pegaso Telematic University); Federico Becattini (University of Siena)*; Marco Bertini (University of Florence)

Truth or Lie: An Audio-Visual Approach to Deception Detection

Antonio Galli (University of Naples Federico II); Michela Gravina (University of Naples Federico II)*; Antonio Elia Pascarella (University of Naples Federico II); Francesco Di Serio (University of Naples Federico II); Daiana Cipollaro (University of Naples Federico II); Vincenzo Moscato (University of Naples Federico II); Carlo Sansone (University of Naples Federico II)

Efficient Attention on Microcontrollers via Linearized Transformer Blocks

Alberto Ancilotto (Fondazione Bruno Kessler)*; Edoardo Castagnini (Fondazione Bruno Kessler); Elisabetta Farella (Fondazione Bruno Kessler)

Few-Shot Anomaly Classification by Learning to Inpaint Nominal Images

Rubin Carkaxhia (University of Bologna); Musawar Ali (University of Bologna)*; Nicola Fioraio (Eyecan.ai); Luigi Di Stefano (University of Bologna); Samuele Salti (University of Bologna)

ExDD: Explicit Dual Distribution Learning for Surface Defect Detection via Diffusion Synthesis

Muhammad Aqeel (University of Verona)*; Federico Leonardi (University of Verona); Francesco Setti (University of Verona)

Improving generative data augmentation with prior-knowledge for dysplasia grading of colorectal polyps

Roberto Craparotta (University of Turin); Desislav Ivanov (University of Turin); Carlo Alberto Barbano (University of Turin; Télécom Paris)*; Enzo Tartaglione (Télécom Paris); Alessandro Gambella (University of Turin); Luca Cavallo (University of Turin); Paola Cassoni (University of Turin); Luca Bertero (University of Turin); Marco Grangetto (University of Turin)

High-Fidelity Infrared Scenario Simulation with Conditional CycleGAN

Margherita Barbuti (Leonardo S.p.A.)*; Fabio Dell'Acqua (Università di Pavia); Roberto Conte (Leonardo S.p.a.); Simone Finelli (Leonardo S.p.a.)

Automated Instance Label Generation for Traffic Domain to Enhance YOLACT-YOLOv9

Dick Scholte (Eindhoven University of Technology)*; Matthijs Zwemer (ViNotion B.V.); Peter de With (Eindhoven University of Technology); Egor Bondarev (Eindhoven University of Technology)

Zero-Shot Fibers Detection in Scanning Electron Micrographs

Pierluigi Carcagni' (National Research Council of Italy - Institute of Applied Sciences & Intelligent Systems)*; Marco Del Coco (National Research Council of Italy - Institute of Applied Sciences & Intelligent Systems); Gennaro Gentile (National Research Council of Italy - Institute for Polymers, Composites and Biomaterials); Maria Cristina Cocca (National Research Council of Italy - Institute for Polymers, Composites and Biomaterials); Melania Paturzo (National Research Council of Italy - Institute of Applied Sciences & Intelligent Systems); Marco Leo (National Research Council of Italy - Institute of Applied Sciences & Intelligent Systems)

Augmented Reality in Cultural Heritage: A Dual-Model Pipeline for 3D Artwork Reconstruction

Daniele Pannone (Sapienza University of Rome)*; Alessia Castronovo (Sapienza University of Rome); Maurizio Mancini (Sapienza University of Rome); Gian Luca Foresti (University of Udine); Claudio Piciarelli (University of Udine); Rossana Gabrieli (Sapienza University of Rome); Muhammand Yasir Bilal (Sapienza University of Rome); Danilo Avola (Sapienza University of Rome)

UAV-based photovoltaic farm maintenance system

Michał Chlebiej (Nicolaus Copernicus University in Toruń); Tomasz Grzona (Nicolaus Copernicus University in Toruń)*; Jakub Kluczewski (Nicolaus Copernicus University in Toruń); Andrzej Rutkowski (Nicolaus Copernicus University in Toruń)

An Investigation on Incremental Learning from Unbalanced Streamed Data

Guido Borghi (University of Modena and Reggio Emilia)*; Gabriele Graffieti (University of Bologna); Roberto Vezzani (University of Modena and Reggio Emilia)

MECAD :A multi-expert architecture for continual anomaly detection

Malihe Dahmardeh (University of Verona)*; Francesco Setti (University of Verona)

RAW-Mix: Region-Aware Mixing for Unsupervised Domain Adaptation in Object Detection

Andualem Welabo Tulu (University of Trento)*; Nicola Conci (University of Trento)

Enhancing Scene Segmentation Using Spatial Relationships: A Case Study on Byzantine Seals

Ege Şendoğan (Sorbonne Université)*; Victoria Eyharabide (Sorbonne Université); Isabelle Bloch (Sorbonne Université)

When Does Pruning Benefit Vision Representations?

Enrico Cassano (Università Degli Studi di Torino)*; Riccardo Renzulli (Università Degli Studi di Torino); Andrea Bragagnolo (Links Foundation); Marco Grangetto (Università Degli Studi di Torino)

KairosAD: A SAM-Based Model for Industrial Anomaly Detection on Embedded Devices

Uzair Khan (university of Verona)*; Franco Fummi (University of Verona); Luigi Capogrosso (University of Verona)

Efficient Odd-One-Out Anomaly Detection

Silvio Chito (Politecnico di Torino); Paolo Rabino (Politecnico di Torino)*; Tatiana Tommasi (Politecnico di Torino)

Segmenting Anything in Architecture: a Tailored Approach to Segmenting Floor Plan Documents

Claudio Piciarelli (Università degli Studi di Udine); Silvia Zottin (Università degli Studi di Udine)*; axel de nardin (Università degli Studi di Udine); gian luca foresti (Università degli Studi di Udine)

WARD: Weather-Aware Road Surface Condition Monitoring Dataset

Soňa Drocárová (The Czech Academy of Sciences, Institute of Information Theory and Automation); Tomáš Kerepecký (The Czech Academy of Sciences, Institute of Information Theory and Automation); Adam Novozámský (The Czech Academy of Sciences,  Institute of Information Theory and Automation)*

Friday, 19/9

Keynote 3 - Learning to see the hidden world: a perspective on casual representations

Invited speaker: Francesco Locatello

Abstract: Deciphering observations into structural knowledge of the world is a key component of the scientific discovery process and a longstanding challenge for AI. In this talk, I will follow the methodological trajectory from sparse coding and ICA to the deep learning era of disentangled representations of scenes, objects, and their causal relations. Alongside the methodologies, I will discuss the theoretical identifiability challenges and how they have driven method development until their recent (partial) solution. Finally, I will discuss the development of the application side, first within machine learning and now also to other scientific disciplines.

Oral session 3.A - Deep Learning

10:30 - 10:50

CoLoR-GAN: Continual Few-Shot Learning with Low-Rank Adaptation in Generative Adversarial Networks

Munsif Ali (University of Parma); Leonardo Rossi (University of Parma)*; Massimo Bertozzi (University of Parma)

10:50 - 11:10

A Second-Order Perspective on Pruning at Initialization and Knowledge Transfer

Leonardo Iurada (Politecnico di Torino)*; Beatrice Occhiena (Politecnico di Torino); Tatiana Tommasi (Politecnico di Torino)

11:10 - 11:30

How to Train Your Metamorphic Deep Neural Network

Thomas Sommariva (Università degli studi di Modena e Reggio Emilia)*; Simone Calderara (Università degli studi di Modena e Reggio Emilia); Angelo Porrello (Università degli studi di Modena e Reggio Emilia)

11:30 - 11:50

Confidently Biased (ConB): a per-sample confidence approach for unsupervised model debiasing

Massimiliano Ciranni (MaLGa, DIBRIS, University of Genoa); Aliaksandr Shcharbitski (University of Genoa); Vittorio Murino (Istituto Italiano di Tecnologia); Vito Paolo Pastore (MaLGa, DIBRIS, University of Genoa)*

11:50 - 12:10

There is No Model to Beat Them All: Recommendations for Deep Learning Model Selection when Training on Synthetic Images

Joachim Rüter (German Aerospace Center (DLR))*; Johann Dauer (German Aerospace Center (DLR)); Umut Durak (German Aerospace Center (DLR))

Oral session 3.B - Applications

15:30 - 15:50

Image Specific Protection against Image Manipulation

Filippo Bartolucci (University of Bologna)*; Giuseppe Lisanti (University of Bologna)

15:50 - 16:10

A Reinforcement Learning Framework for Optimizing Predictive Maintenance of Photovoltaic Systems

Andres Felipe Perez Murcia (University of Manitoba)*; Giulia Lombardi (University of Trento)

16:10 - 16:30

Incremental Object-Based Novelty Detection with Feedback Loop

Simone Caldarella (University of Trento)*; Elisa Ricci (University of Trento, Fondazione Bruno Kessler); Rahaf Aljundi (Toyota Motor Europe)

16:30 - 16:50

RoadFusion: Latent Diffusion Model for Pavement Defect Detection

Muhammad Aqeel (University of Verona)*; Kidus Dagnaw Bellete (University of Verona); Francesco Setti (University of Verona)

16:50 - 17:10

Intrinsic Training Signals for Federated Learning Aggregation

Cosimo Fiorini (University of Modena and ReggioEmilia)*; Matteo Mosconi (University of Modena and ReggioEmilia); Pietro Buzzega (University of Modena and ReggioEmilia); Riccardo Salami (University of Modena and ReggioEmilia); Simone Calderara (University of Modena and ReggioEmilia)

Poster session 3

Diffusion-Detect: synthetic and real data towards a robust Fire and Smoke detection

Dadan Khan (University of Genoa)*; Jacopo Motta (Aitek S.p.A., Genoa); Francesca Odone (University of Genoa)

Estimation of particle growth kinetics via Physics-Informed Neural Networks

Usman Syed (University of Verona); Eros Radicchi (University of Verona); Paolo Marone (Istituto Internazionale del Marmo); Francesco Enrichi (University of Verona); Adolfo Speghini (University of Verona); Federico Cunico (University of Verona)*

Multi-Swimmer Drowning Detection Using a Custom Annotated Underwater Dataset and Real-Time AI

Sarah Kohail (Zayed University)*; Hamad Alzaabi (Zayed University); Saif Alzaabi (Zayed University)

Unravelling Neurodivergent Gaze Behaviour through Visual Attention Causal Graphs

Giuseppe Cartella (University of Modena and Reggio Emilia)*; Vittorio Cuculo (University of Modena and Reggio Emilia); Alessandro D'Amelio (University of Milan); Rita Cucchiara (University of Modena and Reggio Emilia); Giuseppe Boccignone (University of Milan)

Evaluating Attribute Confusion in Fashion Text-to-Image Generation

Ziyue Liu (University of Verona)*; Federico Girella (University of Verona); Yiming Wang (Fondazione Bruno Kessler); Davide Talon (Fondazione Bruno Kessler)

NeSyLAD: A Neuro-Symbolic Approach for Unsupervised Logical Anomaly Detection

Malihe Dahmardeh (University of Verona)*; Mohsen Saadatpour Moghaddam  (University of Verona); Francesco Manigrasso (University Of Politecnico Di Torino); Lia Morra (University of Politecnico Di Torino); Francesco Setti (University of Verona)

Multimodal Deepfake Detection with Large Vision-Language Models: the State of the Art

Awais Abdul Khaliq (Università degli Studi di Milano)*; Stefano Montanelli (Università degli Studi di Milano)

HoSVD-NSST Framework for Secure Dual Medical Image Watermarking using Deep Learning

Roopkamal K. Thind (Thapar Institute of Engineering and Technology, Patiala, Punjab); Rajat Sood (Thapar Institute of Engineering and Technology, Patiala, Punjab); Jyoti  Rani (Manipal University, Jaipur, India); Ashima Anand (Thapar Institute of Engineering and Technology, Patiala, Punjab); Anamika Sharma (Thapar Institute of Engineering and Technology, Patiala, Punjab)*

Single-Path Precision: Differentiable Bit-Width & Numeric-Format Learning for FPGA-Efficient Neural Networks

Emanuel Trabes (University of Mons)*; Kawther Dellel (University of Mons); Carlos Valderrama (University of Mons); Abdelmalik Taleb-Ahmed (Université Polytechnique Hauts-de-France)

Diversified in-domain synthesis with efficient fine-tuning for few-shot classification

Nicola Dall'Asen (University of Trento)*; Victor Turrisi (Apple); Niculae Sebe (University of Trento); Yiming Wang (Fondazione Bruno Kessler); Elisa Ricci (University of Trento)

Classification of gliomas from whole slide imaging based on lymphocyte pattern analysis

Aravindan Arun Nadaradjane (LIP6)*; Julie Lerond (Institut du Cerveau - Paris Brain Institute, CNRS); Mehdi  Touat (Department of Neuro-oncology, Sorbonne University, CNRS, Paris Brain Institute, La Pitié Salpêtrière - Charles Foix Hospital Group); Franck Bielle (Sorbonne University, Paris Brain Institute, La Pitié Salpêtrière - Charles Foix Hospital Group, Department of Neuropathology); Isabelle Bloch (Sorbonne Université, CNRS, LIP6)

MORE: A framework for stable white blood cell morphological classification and report generation

Luca Zedda (University of Cagliari)*; Andrea Loddo (University of Cagliari); Cecilia Di Ruberto (University of Cagliari)

ReMAR-DS: Recalibrated Feature Learning for Metal Artifact Reduction and CT Domain Transformation

Mubashara Rehman (University of Udine)*; Niki  Martinel (University of Udine); Michele  Avanzo (Centro di Riferimento Oncologico di Aviano IRCCS); Riccardo  Spizzo (Centro di Riferimento Oncologico di Aviano IRCCS); Christian  Micheloni (University of Udine)

Graph-based Evaluation of Visual Brain Decoding from fMRI Data

Mohammad Moradi (University of Catania)*; Morteza Moradi (University of Catania); Marco Grassia (University of Catania); Giuseppe  Mangioni (University of Catania)

Efficient Single Image Super-Resolution for Images with Spatially Varying Degradations 

Simone Bianco (University of Milano - Bicocca)*; Luca Cogo (University of Milano - Bicocca); Gianmarco Corti (University of Milano - Bicocca); Raimondo Schettini (University of Milano - Bicocca)

AI-driven visual monitoring of industrial assembly tasks

Mattia Nardon (Fondazione Bruno Kessler); Stefano Messelodi (Fondazione Bruno Kessler); Antonio Granata (Meccanica del Sarca); Fabio Poiesi (Fondazione Bruno Kessler); Alberto Danese (Meccanica del Sarca); Davide Boscaini (Fondazione Bruno Kessler)*

Segment Anything for Satellite Imagery: A Strong Baseline and a Regional Dataset for Automatic Field Delineation

Carmelo Scribano (Università di Modena e Reggio Emilia)*; Elena Govi (Università di Modena e Reggio Emilia); Paolo Bertellini (Abaco Group); Simone Parisi (Abaco Group); Giorgia Franchini (Università di Modena e Reggio Emilia); Marko Bertogna (Università di Modena e Reggio Emilia)

A Handful of Data: Evaluating Few–shot Incremental Landmark Detection

Nikil Patel (University of Oxford)*; Allison Clement (University of Oxford); Julian Wyatt ( University of Oxford); Irina Voiculescu (University of Oxford); Roberto Divia (University of Genoa); Davide Marinelli (University of Genoa); Massimiliano Ciranni (University of Genoa); Vito Pastore (University of Genoa)

Pre-NeRF: Evaluating Preprocessing Approaches to Mitigate Real-World Corruptions in NeRF Reconstruction

Alexander Suen (Stanford University)*; Cayden Gu (Stanford University)

Skeleton-based action recognition for the biomechanical risk condition assessment

Irene Gennarelli (University of Udine); Gianluca De Nardi (University of Udine); Alberto Ranavolo (INAIL); Christian Micheloni (University of Udine); Niki Martinel (University of Udine)*

T-EVO: Tracking in Egovision for Online Visual Episodic Memory

Zaira Manigrasso (Università di Udine)*; Antonio Finocchiaro (Università di Catania); Davide  Manara (Università di Udine); Rosario Forte (Università di Catania); Moritz Nottebaum (Università di Udine); Matteo Dunnhofer (Università di Udine); Giovanni Maria Farinella (Università di Catania); Antonino Furnari (Università di Catania); Christian Micheloni (Università di Udine)

Historical postcards classification combining visual content and text description

Matthieu Pelingre (Université de Lorraine); Salvatore Tabbone (Loria)*

Multimodal Emotion Recognition via Multilevel Fusion of Visual, Audio, and Textual Data

José Salas-Cáceres (ULPGC)*; Javier Lorenzo-Navarro (ULPGC); Modesto Castrillón-Santana (ULPGC)

Momaku: A Retinal Image Annotation Platform

Michele Cannito (Politecnico di Torino)*; Anna Maria Llopart (Universitat Pompeu Fabra); Nerea Ferrara (Universitat Pompeu Fabra); Massimo Salvi (Politecnico di Torino); Alicia Serra (Hospital Sant Joan de Deu); Arnau Valls (Hospital Sant Joan de Deu); Oscar Camara (Universitat Pompeu Fabra); Adrian Galdran (Universitat Pompeu Fabra)

Multimodal Distillation for Video-based Sleep Behavior Analysis

Jacopo Donà (University of Verona)*; Marco Carletti (EVS Embedded Vision Systems SRL); Davide Tonon (Ospedale Sacro Cuore Don Calabria); Gianluca Rossato (Ospedale Sacro Cuore Don Calabria); Vittorio Murino (Istituto Italiano di Tecnologia, University of Verona)

The conference will take place fully in person, with no online streaming. All presentations must be delivered on-site.

Oral presentations

• Each oral presentation is allocated 15 minutes, and 5 minutes are reserved for questions, speaker transitions, and introductions by the session chair. Total time on the stage: 20 minutes.

• Given the tight schedule, this time limit will be strictly enforced. Speakers are required to leave the podium once their time is over.

• Each speaker will present using their own laptop, and the connection will be via HDMI.

• Technical instructions will be provided soon.

Poster presentation

• To be announced.