Welcome to AImageLab
Research at AImageLab covers topics of Computer Vision, Pattern Recognition & Machine Learning, Artificial Intelligence, and Multimedia applied to optical images and videos as well as data from different sensors. In addition, AImageLab is currently involved in research activities on vision-based human-computer interaction.
The main research projects concern: Video surveillance, Machine vision and robot vision, Medical imaging, Human-centered Multimedia, Content-based retrieval, People detection and tracking, Human behaviour understanding, Egocentric vision and Embedded sensors.
Open Research Positions @ AImageLab · 27 Jul 2020
Seminar "Autonomous Driving: the missing piece" · 26 Oct 2020
On October, 29th at 9.00, Alberto Broggi (VisLab srl) will give an invited talk on "Autonomous Driving: the missing piece". The event is organized as the first of a cycle of seminars on AI for Automotive, as part of the "Neural Network Computing, AI and Machihne Learning for Automotive" and "Learning Algorithms for Smart Intelligent Systems" courses (Prof. Cucchiara, Baraldi) and will be streamed online.
Autonomous driving is currently a very hot topic, and the whole automotive industry is now working hard to transition from research to products. Deep learning and the progress of silicon technology are the main enabling factors that boosted the interest of industry and are currently pushing the automotive sector towards self-driving cars. Computer vision is one of the most important sensing technologies thanks to the extremely dense data set it can provide as output: object classification and recognition, precise localization, 3D reconstruction are just examples. This presentation addresses the missing piece that will allow future vehicles to make a full and efficient use of computer vision.
See the announcement on the AI Academy website for the link to the on-line streaming.
Rita Cucchiara at Eredità delle Donne Festival · 24 Oct 2020
EREDITÀ DELLE DONNE FESTIVAL
Scientists, economists, businesswomen, politicians, writers and artists confronted for three days on how to face the post-Covid-19 world: a "new" world, with its challenges, its suffering, its potential. This is Eredità delle Donne Festival.
Rita Cucchiara is going to give her vision on Saturday 24th at 5:20 pm in the panel Letters from the Future.
Intelligenza Artificiale, quale etica per il futuro. La via italiana ed europea all'innovazione. · 16 Oct 2020
quale etica per il futuro.
La via italiana ed europea all'innovazione.
Lo stato dell’arte dell'AI, le sue applicazioni più promettenti, l’impatto su società, economia, politica e mondo del lavoro, la via italiana ed europea all’IA e la lotta alle malattie.
Questi sono solo alcuni dei grandi temi che verranno affrontati dalla Prof. Rita Cucchiara e da numerosi altri esponenti del mondo dell'innovazione digitale, guidati da Alessio Jacona, martedì 20 Ottobre, ore 15 in streaming su ANSA.it
➡ Review the event: https://vimeo.com/470089348
Paper Accepted for publication on T-PAMI · 14 Oct 2020
"Warp and Learn: Novel Views Generation for Vehicles and Other Objects"
by A.Palazzi, L.Bergamini, S.Calderara and R. Cucchiara
has been accepted for publication on IEEE Transactions on Pattern Analysis and Machine Intelligence
Paper accepted at NeurIPS 2020 · 13 Oct 2020
"Dark Experience for General Continual Learning: a Strong, Simple Baseline"
by P. Buzzega, M. Boschini, A. Porrello, D. Abati and S. Calderara
has been accepted (poster) at NeurIPS 2020, 6-12 December 2020, Vancouver, Canada.
'About Time' - Seminar by Prof. Arnold Smeulders · 25 Sep 2020
'About Time' - Seminar by Prof. Arnold Smeulders
Equivariant, temporal representations and networks are the topics that will be discussed during the seminar held by Prof. Arnold Smeulders, guest at the Tecnopolo of Modena.
Wednesday 30th September, 4:00 p.m.
Tecnopolo Event Room, 2, Vivarelli street, 41125 Modena
Modena Smart Life 2020 Intelligenza Artificiale, curiosità e immaginazione · 24 Sep 2020
Modena Smart Life 2020 Intelligenza Artificiale, curiosità e immaginazione
Il 25 Settembre dalle 11:00 alle 11:15 Rita Cucchiara e Lorenzo Baraldi saranno collegati il live streaming per parlare di IA e immaginazione in occasione del Modena Smart Life, il Festival della Cultura Digitale, giunto alla sua quinta edizione.
Consulta il link per maggiori informazioni:
MACHINE LEARNING E DEEP LEARNING - Corso teorico e pratico · 23 Sep 2020
Fondazione Democenter, in partenariato con la AI Academy del Dipartimento di Ingegneria Enzo Ferrari e il laboratorio di Ingegneria Informatica AImageLab ripropone una nuova edizione del corso in Machine Learning e Deep Learning con l’obiettivo di fornire a figure professionali con un background informatico competenze specialistiche aggiuntive per la progettazione e lo sviluppo di algoritmi di machine learning, per la raccolta, l’elaborazione, analisi di dati e l’identificazione di modelli.
Il corso, in accordo con i docenti, sarà erogato ON LINE in modalità sincrona (in diretta) prevedendo 2 sessioni di lezioni a settimana della durata ciascuna di 4 ore.
Per maggiori informazioni clicca qui.
Workshop - Towards European Sovereignty and Autonomy in AI · 17 Sep 2020
The European Research and Innovation Days is the European Commission’s annual flagship Research and Innovation event, bringing together policymakers, researchers, entrepreneurs and the public to debate and shape the future of research and innovation in Europe and beyond.
Prof. Rita Cucchiara is invited to discuss on how the different European initiatives will interact and boost their own resources to establish the basis for the European sovereignty and autonomy in AI in a special panel composed by Ioannis Kompatsiaris, Bernhard Schölkopf and Nuria Oliver, during the workshop named "Towards European Sovereignty and Autonomy in AI".
UNIMORE@ELLIS · 15 Sep 2020
We are proud to announce that AImageLab and the Interdepartmental Artificial Intelligence Research and Innovation Center of UNIMORE become part of the European network ELLIS!!!
The European Lab for Learning and Intelligent Systems (ELLIS) is an intergovernmental organization comprising 30 laboratories (ELLIS Units) in 14 European countries devoted to research on machine learning and intelligent systems.
Click here for more information.
Inter-Homines: Best Demo Award at ECCV 2020 · 01 Sep 2020
AImageLab receives two ISCRA-C grants · 23 Jul 2020
AImageLab has received two ISCRA-C grants from NVIDIA, for accessing the supercomputing facilities of MARCONI100. MARCONI100 is the new accelerated not conventional Marconi partition, available from April 2020, and equipped with NVIDIA Volta V100GPUs.
Read more about the NVIDIA AI Technology Centre.
Workshop Human-centered Vision: from Body Analysis to Learning and Language · 07 Jul 2020
Giovedì 9 luglio 2020 - 15.00 - 17.00
15.10-15.30: introduzione ai Laboratori di Ricerca
- Biometric System Laboratory, Università di Bologna - campus di Cesena (Prof. Maltoni)
- AImageLab, Università di Modena e Reggio Emilia (Prof.ssa Cucchiara)
15.30 - 15.45: Real-Time Continual Learning from Natural Video Streams (Vincenzo Lomonaco e Lorenzo Pellegrini)
15.45 - 16.00: Classificazione Rehearsal-Based in Continual Learning. (Matteo Boschini, Pietro Buzzega e Simone Calderara)
16.00 - 16.15: People Behavior and Face Understanding (Roberto Vezzani, Stefano Pini, Matteo Fabbri e Fabio Lanzi)
16.15 - 16.30: The challenge of Morphing for border control (Matteo Ferrara e Annalisa Franco)
16.30 - 16.45: Modelli generativi per Image Translation e Continual Learning (Gabriele Graffieti)
16.45 - 17.00: Vision, Language and Action: from Captioning to Embodied AI (Lorenzo Baraldi, Federico Landi e Marcella Cornia)
AImageLab presenting three papers at CVPR 2020 · 17 Jun 2020
AImageLab is presenting three papers at CVPR 2020:
- "Compressed Volumetric Heatmaps for Multi-Person 3D Pose Estimation"
by M. Fabbri, F. Lanzi, S. Alletto, S. Calderara, R. Cucchiara
Date: Wednesday, June 17 & Thursday, June 18
Q&A Time: 12:00-14:00 and 00:00-02:00 (PST) / 21:00-23:00 and 09:00-11:00 (CEST)
Session: Poster 2.2 - Face, Gesture, and Body Pose; Motion and Tracking; Representation Learning
[paper] [code] [video presentation]
- "Conditional Channel Gated Networks for Task-aware Continual Learning"
by D. Abati, J. Tomczak, T. Blankevoort, S. Calderara, R. Cucchiara, B. E. Bejnordi
Date: Tuesday, June 16 & Wednesday, June 17
Q&A Time: 16:00-18:00 and 04:00-06:00 (PST) / 01:00-03:00 and 13:00-15:00 (CEST)
Session: Oral 1.4C - Transfer/Low-Shot/Semi/Unsupervised Learning
[paper] [oral presentation]
- "Meshed-Memory Transformer for Image Captioning"
by M. Cornia, M. Stefanini, L. Baraldi, R. Cucchiara
Date: Thursday, June 18 & Friday, June 19
Q&A Time: 09:00-11:00 and 21:00-23:00 (PST) / 18:00-20:00 and 06:00-08:00 (CEST)
Session: Poster 3.1 - Recognition (Detection, Categorization); Video Analysis and Understanding; Vision + Language
[paper] [code] [video presentation]
See the program for more.
Webinar Applicazioni dell'Intelligenza Artificiale nei settori industriali: le tecnologie Machine learning e Deep learning · 10 Jun 2020
Applicazioni dell’Intelligenza Artificiale nei settori industriali: le tecnologie Machine learning e Deep learning – 25 e 26 Giugno
Si terranno nelle giornate del 25 e 26 Giugno 2020 le due sessioni interattive di formazione previste nel corso “Applicazioni dell’Intelligenza Artificiale nei settori industriali: le tecnologie Machine learning e Deep learning”.
Democenter, in partnership con l'AI Academy di UNIMORE offre la possibilità di dare continuità ad iniziative formative sulla Intelligenza artificiale come parte integrante e fondamentale della trasformazione digitale delle imprese e lo fa proponendo le competenze presenti all’interno del laboratorio di ricerca AImageLab.
Il corso si propone di fornire a Operations Manager, R&D Manager, Responsabili IT, Responsabili delle varie aree aziendali una panoramica aggiornata delle applicazioni del Machine learning e Deep learning, delle problematiche relative a quali software e hardware adottare e quali sono i costi di tali soluzioni.
Ing. Simone Calderara, professore associato di Machine Learning e Deep Learning del Dipartimento di Ingegneria “Enzo Ferrari” dell’Università di Modena e Reggio Emilia, membro senior del laboratorio di computer vision, AI e Machine Learning AImageLab (direttore Prof. Rita Cucchiara) e membro senior dell’ AI Academy UNIMORE. Referente per UNIMORE nel laboratorio nazionale Artificial Intelligence and Intelligent System e nel gruppo di lavoro APRE Digital.
Tutte le informazioni e le modalità di partecipazione al seguente LINK.
Intervista a "Dialoghi sul nostro tempo" de L'Espresso · 29 May 2020
Una approfondita riflessione sulle potenzialità delle tecnologie che si basano sui sistemi di Intelligenza Artificiale, spaziando dalla panoramica sullo stato attuale della ricerca nazionale ed internazionale ad un approfondimento sull'AI a supporto della medicina. In che modo i sistemi intelligenti possono essere d'aiuto per fronteggiare gli inevitabili cambiamenti generati della pandemia da COVID-19? Quale contributo mette in campo l'AI? Cosa si prospetta nel prossimo futuro?
Ne parla la Prof. Rita Cucchiara, Responsabile di AImageLab e Direttore del LN AIIS del CINI, intervistata insieme a Roberto Cingolani, Head of Innovation in Leonardo-Finmeccanica ed ex Direttore Scientifico dell'IIT, da Marco Damilano durante la trasmissione Dialoghi sul Nostro Tempo de L'Espresso.
YNAP AI LAB@UNIMORE · 28 May 2020
As a leading company in the use of new technologies, Yoox Net-A-Porter Group, the most advanced fashion hub in Italy for Artificial Intelligence and Visual Recognition, has decided to join forces with the University of Modena and Reggio Emilia and has created a joint research program dedicated to Artificial Intelligence and Computer Vision in the field of fashion.
Press release available here.
NVAITC Webinar for the CINI community · 05 May 2020
To present the NVIDIA AI Nation program and the collaboration between NVIDIA AI Technology Center and the Italian research community, which has been launched in Modena last January, a webinar will be held on May 6th from 11:30 to 13:00. During the webinar, the Director of the CINI AIIS Lab, together with NVIDIA and CINECA will present the scope of the program and the computational resources available in Italy through CINECA.
Agenda of the webinar:
Introduction – Rita Cucchiara (UNIMORE)
NVAITC program, how to apply – Frederic Pariente, Giuseppe Fiameni (NVIDIA)
How to access CINECA resources – Marco Rorro (CINECA)
Registration is available at the following link.
Galleria Estense of Modena: Walking through the museum, while staying home · 08 Apr 2020
The Galleria Estense of Modena, in partnership with AImageLab, organizes virtual tours that let visitors walk through the halls of the art gallery without leaving their homes.
In particular, connected users will be able to participate a virtual tour guided by a museum operator: these visits allow the continuation of the didactic proposal reserved to schools and the usual Sunday visits open to everyone. Both the activities were interrupted by the Covid-19 health emergency.
The 3D reconstruction of the Galleria Estense allows participants to move freely between the exhibition spaces and to communicate verbally with the guide, who can answer questions but also share analytical insights of the works and other documents related to the collections on display. This innovative experience in terms of involvement and cultural deepening is made possible thanks to Matterport technology.
The 3D acquisition of Galleria Estense was carried on by AImageLab as a part of a larger research project on Cultural Heritage and Digital Humanities supported by Fondazione di Modena.
We are proud that our work may now serve as a tool to promote education and culture in a time of emergency.
High Skills for Research and Technological Transfer · 03 Apr 2020
Our research project "Automated detection of the Inferior Alveolar nerve Canal (IAC) in Cone Beam Computed Tomography (CBCT) using deep-learning techniques" has been selected by Emilia Romagna region in the call Alte Competenze per la Ricerca e il Trasferimento Tecnologico (High Skills for Research and Technological Transfer). Together with the research team of Prof. Alexandre Anesi, AImageLab will receive a 30,000.00 € grant for a research fellow position.
Accepted Paper at ICRA 2020 · 02 Mar 2020
"SMArT: Training Shallow Memory-aware Transformers for Robotic Explainability"
by M. Cornia, L. Baraldi, R. Cucchiara
has been accepted to ICRA 2020, that will take place from May 31st to June 4th in Paris, France.
MORE AI - NVIDIA AI Nation, CINI ed UNIMORE · 07 Jan 2020
Il 16 gennaio dalle ore 14:30, presso il Tecnopolo di Modena, si terrà una giornata di presentazione dell'iniziativa "AI Nation NVIDIA" e del nascituro Laboratorio Congiunto NVAITC-UNIMORE (NVIDIA Artificial Intelligence Technology Center). L'evento è organizzato in collaborazione con il Laboratorio Nazionale CINI su Artificial Intelligence and Intelligent Systems (AIIS) e vedrà la partecipazione di UNIMORE, CINI, NVIDIA e aziende del territorio.
Il programma dell'evento e il link per la registrazione sono disponibili sul sito dell'AI Academy di Unimore.
Rita Cucchiara's keynote at the Georgetown University · 05 Nov 2019
Rita Cucchiara is giving an invited keynote on "Visual Intelligence: Research and Applications for Human-centered AI" at the Georgetown University in Washington, DC. The seminar is sponsored by the Georgetown University Italian Research Institute of Georgetown College, in collaboration with the Embassy of Italy, the Italian Cultural Institute, and the Georgetown University Center for Security and Emerging Technology.
Slides are available.
Abstract: Research and Applications for Human-centered AI Over the past decade, the joined research in Machine Learning and Computer Vision achieved impressive results worldwide, as of the most successful area of Artificial Intelligence. The capability of understanding images and video content reached new solutions both for autonomous intelligent systems and for an augmenting intelligence of human activities. The talk will focus on some aspects of what is called Visual Intelligence for AI-based systems, which mimics the typical human ability to understand the world, predict events, and imagine the possible through the visual perception. In particular, new research results for salient object recognition, attentional analysis and details detection in images will be presented, with reference of some projects carried out at AImagelab, in Italy, together with national and international companies. This visual intelligence capability is directly applicable in different contexts: the prediction of car-driver attention, the robot exploration of unknown environments, the detection of humans with their actions, the automatic description of image content by language as well as a novel way of knowledge extraction from cultural heritage data. Imagination, creativity, and curiosity-driven behaviour can be transferred to machines, now, and this will be potentially the core of future generations of human-centered AI.
Link to the event page
Paper accepted in Transactions on Image Processing · 30 Sep 2019
Our paper Spaghetti Labeling: Directed Acyclic Graphs for Block-Based Connected Components Labeling has been accepted in Transactions on Image Processing.
See the paper and the GitHub repository with the source code.
Connected Components Labeling is an essential step of many Image Processing and Computer Vision tasks. Since the first proposal of a labeling algorithm, which dates back to the sixties, many approaches have optimized the computational load needed to label an image. In particular, the use of decision forests and state prediction have recently appeared as valuable strategies to improve performance. However, due to the overhead of the manual construction of prediction states and the size of the resulting machine code, the application of these strategies has been restricted to small masks, thus ignoring the benefit of using a block-based approach. In this paper, we combine a block-based mask with state prediction and code compression: the resulting algorithm is modeled as a Directed Rooted Acyclic Graph with multiple entry points, which is automatically generated without manual intervention. When tested on synthetic and real datasets, in comparison with optimized implementations of state-of-the-art algorithms, the proposed approach shows superior performance, surpassing the results obtained by all compared approaches in all settings.
Best Paper Award at CAIP 2019 · 27 Sep 2019
Our paper How does Connected Components Labeling with Decision Trees perform on GPUs has been awarded as the best paper of the 18th International Conference on Computer Analysis of Images and Patterns.
See the paper and the GitHub repository with the source code.
In this paper the problem of Connected Components Labeling (CCL) in binary images using Graphic Processing Units (GPUs) is tackled by a different perspective. In the last decade, many novel algorithms have been released, specifically designed for GPUs. Because CCL literature concerning sequential algorithms is very rich, and includes many efficient solutions, designers of parallel algorithms were often inspired by techniques that had already proved successful in a sequential environment, such as the Union-Find paradigm for solving equivalences between provisional labels. However, the use of decision trees to minimize memory accesses, which is one of the main feature of the best performing sequential algorithms, was never taken into account when designing parallel CCL solutions. In fact, branches in the code tend to cause thread divergence, which usually leads to inefficiency. Anyway, this consideration does not necessarily apply to every possible scenario. Are we sure that the advantages of decision trees do not compensate for the cost of thread divergence? In order to answer this question, we chose three well-known sequential CCL algorithms, which employ decision trees as the cornerstone of their strategy, and we built a data-parallel version of each of them. Experimental tests on real case datasets show that, in most cases, these solutions outperform state-of-the-art algorithms, thus demonstrating the effectiveness of decision trees also in a parallel environment.
Third Place at the 2019 ISIC Challenge · 26 Sep 2019
Modena Smart Life 2019 · 24 Sep 2019
Human society and artificial intelligence: this is the subject of the 4th edition of Modena Smart Life, the Festival of digital culture located in Modena from 27th to 29th September 2019.
The technological evolution and the introduction of new intelligent systems and tools not only pervade communications, transports, robotics, medical and economic sectors but they suggest to investigate the ever closer human-machine interactions dynamics.
The activities set during the Festival of digital culture explore these aspects.
The UNIMORE AImageLab Professors are called to introduce their studies on the following matters:
Artificial and human intelligence: the engine of business Conference by CNA with Simone Calderara
Artificial intelligence vs humans Lectio Magistralis by Rita Cucchiara
The technological frontier: at what point are we? Keynotes by S. Calderara, L. Baraldi
ARTIficio: Artificial Intelligence meets Art Performative Installation by Rita Cucchiara, Massimiliano Corsini, Lorenzo Baraldi and Marcella Cornia
How the AI see human and his behaviour Keynote by R. Cucchiara
Robot doctor or doctor robot? Artificial Intelligence in medical sector Federico Cabitza in a dialogue with Simone Calderara
PersonArt - Who is your double? · 14 Sep 2019
PersonArt is an interactive system based on Artificial Intelligence that lets you find similarities between real faces and artistic images. The visitor is involved in a course that allows his/her to discover his/her own kind among the portraits of the Galleria Estense di Modena, starting an original and engaging tour. Behind the outcome there are some algorithms based on Deep Neural Networks aiming to identify the face and to extract its salient features in order to calculate similarities between visitor’s face and those on display into the Galleria.
Moreover, PersonArt uses artistic technique in order to decrease the domain gap between photos and artistic pictures, guaranteeing an accurate result transcending the differencesces of texture. The results are elaborated remotely by the high performance server at the seat of the Lab. in the Engineering Department.
Marcella Cornia e Lorenzo Baraldi have introduced these techniques in the most important international meeting in Artificial Vision (CVPR).
You can find the photo box in The Galleria Estense, fourth floor, Largo Porta Sant'Agostino, 337,41121 Modena (MO) Italia during the FestivalFilosofia (13 14 15 September 2019) and Modena Smart Life event (27 28 29 September 2019).
…Who is your double?
LAMV is being used at Facebook to detect harmful content · 03 Aug 2019
A solution for matching and detecting copied videos, developed by AImageLab and Facebook AI Research, is now being used in production scale at Facebook to detect harmful content.
Known as LAMV or TMK (Temporal Matching Kernel), this video-matching technology has been developed by research scientists at Facebook AI Research and by L. Baraldi and R. Cucchiara as part of a collaboration between AImageLab and FAIR. The algorithm is capable of producing a compact and temporal-aware signature of a video, which can be used for video copy detection and video retrieval. The solution combines match kernels, Fourier Transforms and end-to-end learning.
Accepted Oral at BMVC 2019 · 05 Jul 2019
"Embodied Vision-and-Language Navigation with Dynamic Convolutional Filters"
by F. Landi, L. Baraldi, M. Corsini, R. Cucchiara
has been accepted as oral to BMVC 2019, that will take place from September 9th to 12th in Cardiff, UK. The preprint is available.
Seminar announcement · 01 Jul 2019
On July, 15th at 11.00, Iuri Frosio (NVIDIA) will give an invited talk on "Computational Aspects of Deep Reinforcement Learning". The event will be hosted in the Auditorium of the Modena Technopole.
Abstract Deep reinforcement learning has recently emerged as a viable solution for many complex, real world problems. Nonetheless, the development of new deep reinforcement learning algorithms is still badly affected by the numerous instabilities and the long experiment turnaround time. In this talk I will first show how A3C, a standard deep reinforcement learning algorithm, can be accelerated through the adoption of a GPU for inference and training. I will also show that the limited CPU capability for the simulation of a large number of parallel environments as well as the limited CPU-GPU communication bandwidth constitute two significant bottlenecks in this approach. I will finally illustrate our CUDA Learning Environment (CuLE), which allows emulating thousands of Atari games in parallel on the GPU, removing the previously mentioned bottlenecks and opening the door to effective multi-GPU deep reinforcement learning. I will also illustrate few algorithm optimizations that are needed to leverage at best the large amount of data generated by CuLE.
Speaker Bio Iuri Frosio (http://research.nvidia.com/person/iuri-frosio) got his PhD in biomedical engineering at the Politecnico of Milan in 2006. He worked as research fellow at the Computer Science Department of the University of Milan from 2003 and an assistant professor in the same department from 2006 to 2013. In the same period, he worked as a consultant for various companies in Italy and in the US. He joined NVIDIA in 2014 as senior research scientist. His research interests include image processing, computer vision, robotics, parallel programming, machine learning, and reinforcement learning.
AImageLab hosts the first NVIDIA Inception Day in Italy · 24 Jun 2019
On July 8th, NVIDIA Inception and AImageLab are teaming up to organize a meetup for Machine and Deep Learning startups in Modena. The objectives of the event are to present NVIDIA platform solutions for ML/DL, let program members showcase their technology/products and network with their peers, as well as NVIDIA and AImageLab teams.
The meetup will feature advanced Inception members offering deep dives into their DL solutions, a pitching session with a Titan RTX GPU as the prize, a demo running on NVIDIA's DGX supercomputer and an informal networking session.
Obtaining a registration ticket is strongly encouraged. The registration page can be found at this link.
See also the event page on the AI Academy website.
Show, Control and Tell presented at the Workshop on Language and Vision · 16 Jun 2019
Deep Learning for Computer Vision Applications · 23 May 2019
Lorenzo Baraldi, Guido Borghi e Marcella Cornia will give an invited talk at the Department of Information Engineering, Università Politecnica delle Marche.
For further information plase visit the website.
Best Paper Award @ U3DRM 2019 · 13 May 2019
The paper SEMANTIC SEGMENTATION OF BENTHIC COMMUNITIES FROM ORTHO-MOSAIC MAPS - G. Pavoni, M. Corsini, M. Callieri, M. Palma, and R. Scopigno presented at the 2nd International Conference on "Underwater 3D Recording and Modelling - A Tool for Modern Applications and CH Recording" (ISPRS Archives - Volume XLII-2/W10), has been considered among the best papers at the conference.
Conferenza stampa di presentazione della AI Academy · 13 May 2019
Modena e l'Università degli Studi di Modena e Reggio Emilia nei prossimi mesi si andranno a caratterizzare come uno dei più importanti centri di ricerca sull'Intelligenza Artificiale (AI) a livello italiano ed internazionale. Grazie ad una serie di significativi finanziamenti giunti dalla Regione Emilia Romagna, dalla Fondazione Cassa di Risparmio di Modena, cui si aggiunge un finanziamento del Ministero dell'Istruzione, dell'Università e della Ricerca su un progetto relativo alle capacità predittive delle reti neurali, nei prossimi mesi prenderà corpo la nascita di una AI Academy, che potrà disporre di una nuova ala in continuità con l'edificio del Tecnopolo di Modena.
Obiettivi, finalità e costruzione di questo Centro di Ricerca internazionale in Visione ed Intelligenza Artificiale e della AI Academy saranno illustrati in una conferenza stampa convocata per martedì 14 maggio alle ore 11.30.
Maggiori dettagli sono disponibili sul sito web dell'AI Academy.
Talk on Probabilistic Models by Juan Maronas · 09 May 2019
Juan, a PhD student from the PRHLT Research Center of the Universitat Politècnica de València will present his research on Probabilistic Models and their Calibration.
Three papers accepted at CVPR 2019 · 28 Feb 2019
AImageLab will present three papers at CVPR 2019:
Art2Real: Unfolding the Reality of Artworks via Semantically-Aware Image-to-Image Translation (M. Tomei, M. Cornia, L. Baraldi, R. Cucchiara)
Latent Space Autoregression for Novelty Detection (D. Abati, A. Porrello, S. Calderara, R. Cucchiara)
Show, Control and Tell: A Framework for Generating Grounded and Controllable Captions (M. Cornia, L. Baraldi, R. Cucchiara)
Invited talk at the "AI and Future Society" workshop at the British Embassy · 21 Feb 2019
As Director of the CINI Laboratory in "Artificial Intelligence and Intelligent Systems", Rita Cucchiara is giving an invited talk at the "AI and Future Society" workshop at Villa Wolkonsky in Rome. Slides are available.
Intelligenza Artificiale e Sicurezza Stradale: AImagelab premiato nuovamente a livello nazionale · 20 Feb 2019
La tesi magistrale intitolata "Studio e sperimentazione di sistemi di visione artificiale per il monitoraggio del conducente", sviluppata all'interno del laboratorio AImageLab, ha vinto il premio nazionale ANIA come migliore lavoro dell'area tecnico-ingegneristica riguardante il tema della sicurezza stradale, bissando il risultato del 2017.
Autore della tesi è Elia Frigieri, relatore la Prof.ssa Rita Cucchiara, correlatori il Prof. Roberto Vezzani e il dott. Guido Borghi.
La tesi si inserisce nel più ampio settore di ricerca riguardante l'utilizzo di algoritmi di Visione Artificiale e Deep Learning per il monitoraggio automatico del livello di attenzione del conducente durante l'attività di guida.
Link alla notizia pubblicata dalla Fondazione Ania
Clicca qui per maggiori informazioni riguardo al progetto di ricerca
We are glad to announce that the Master's Thesis "Studio e sperimentazione di sistemi di visione artificiale per il monitoraggio del conducente" developed at AImageLab by Elia Frigieri, Rita Cucchiara, Roberto Vezzani and Guido Borghi has won the national contest about Road Security, promoted by ANIA Foundation.
News link (from the ANIA web site)
Click here for more info about the research project
Convegno di presentazione dei risultati del Progetto di Ricerca e Innovazione - La Città Educante · 16 Jan 2019
A Roma, il 23 gennaio 2019 dalle ore 9.30 presso CNR, Sala Convegni in via dei Marrucini snc, si terrà il Convegno di presentazione dei risultati del Progetto di Ricerca e Innovazione
"La Città Educante". Per partecipare è necessario registrarsi a questo link: http://cittaeducante.azurewebsites.net/
Accepted Paper for PAMI Journal · 05 Dec 2018
"Face-from-Depth for Head Pose Estimation on Depth Images"
by G. Borghi, M. Fabbri, R. Vezzani, S. Calderara, R. Cucchiara
has been accepted to be published in TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (TPAMI) journal.
Invited talk at MBDA · 13 Nov 2018
New Patenting Process at AImagelab · 31 Oct 2018
AImagelab is currently patenting an "Authentication System for Low-Light-Level Environments", in collaboration with the University of Modena and Reggio Emilia.
Inventors: Guido Borghi, Stefano Pini, Filippo Grazioli, Roberto Vezzani, Rita Cucchiara
Priority Number: 102018000008237
Keywords: Face Identification, Biometric Identification, Depth Maps, Infrared Sensors
For a brief description and images please visit the project page.
Invited talk at IROS 2018 · 04 Oct 2018
Rita Cucchiara gave an invited talk, "Computer Vision, AI and Robotics for Visual Intelligence", at the IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2018, which was held in Madrid on October 1-5.
Slides from talk are available.
Terzo incontro all’Università sulla visione industriale · 26 Sep 2018
AImageLab ospiterà il “Terzo incontro all’Università sulla visione industriale”, organizzato da Editoriale Delfino Srl.
L’incontro si terrà Giovedì 4 Ottobre 2018 presso il Dipartimento di Ingegneria Enzo Ferrari, .
A chi si rivolge l’evento
Agli studenti delle lauree triennali e magistrali in Ingegneria, per mettere in contatto i futuri ingegneri con i fornitori di una tecnologia, quella dei sistemi di visione, che avrà sempre di più impiego nella produzione industriale. Dopo le presentazioni delle aziende intervenute, l’incontro prevede ampio spazio per un dibattito con gli studenti. L’evento è comunque aperto a chiunque volesse partecipare. Si richiede la registrazione mediante i link in calce.
Il Progetto Visione Industria è dedicato ai sistemi di visione industriali
Intende ricostruirne la storia, offrire materiale di documentazione, proporre supporti per la formazione, segnalare e mettere in collegamento i principali attori in campo industriale, nella ricerca e nei diversi enti e organizzazioni. Si rivolge a tutti coloro che si occupano, a vari livelli, di organizzazione, di gestione e di analisi dei processi produttivi, ai costruttori di macchine e impianti, ai responsabili della qualità, della logistica e della manutenzione ma anche a ricercatori, studenti, operatori della comunicazione. Il progetto, avviato col sostegno di aziende di primo piano del settore e col supporto di un qualificato Comitato Scientifico, verrà via via integrato e periodicamente aggiornato seguendo gli sviluppi tecnologici e di mercato del settore nei prossimi anni.
Link per iscriversi all’evento (GRATUITO): https://www.eventbrite.it/e/biglietti-terzo-incontro-la-visione-industriale-incontra-gli-ingegneri-nelle-universita-47662062512
Per gli studenti:
AImageLab Presentations @ ECCV 2018 Workshops · 08 Sep 2018
Slides of our presentations at ECCV 2018 Workshops are available.
"What was Monet seeing while painting? Translating artworks to photo-realistic images"
by M. Tomei, L. Baraldi, M. Cornia, R. Cucchiara
Oral presentation @ Workshop on Computer VISion for ART Analysis
"Visual-semantic alignment across domains using a semi-supervised approach"
by A. Carraggi, M. Cornia, L. Baraldi, R. Cucchiara
Spotlight presentation @ Multimodal Learning and Applications Workshop
Rita Cucchiara's keynote for the Maria Petrou Prize @ ICPR 2018 · 26 Aug 2018
Accepted Paper at BMVC 2018 · 09 Aug 2018
"Face Verification from Depth using Privileged Information"
by G. Borghi, S. Pini, F. Grazioli, R. Vezzani, R. Cucchiara
has been accepted to BMVC 2018, that will take place from September 3rd to 6th in Newcastle, England.
Accepted Paper at ECCV 2018 · 08 Aug 2018
"Learning to Detect and Track Visible and Occluded Body Joints in a Virtual World"
by M. Fabbri, F. Lanzi, S. Calderara, A. Palazzi, R. Vezzani and R. Cucchiara
has been accepted to ECCV 2018, that will take place from September 8th to 14th in Munich, Germany.
People Tracking by their Pose: the JTA Dataset · 22 Jun 2018
Rita Cucchiara gave an invited talk, "People Tracking by their Pose: the JTA Dataset", at CVPR 2018, which was held at the Calvin L. Rampton Salt Palace Convention Center the week of June 18-22, 2018 in Salt Lake City, Utah.
JUMP – Evento finale · 31 May 2018
Seminar by Prof. Pascal Fua · 23 May 2018
International Workshop on Computer Vision · 02 May 2018
The International Workshop on Computer Vision will consist of high-quality invited talks by distinguished speakers, panels and discussions about emerging themes in computer vision and related fields. The main goal is to encourage people to talk about work in progress, so there will be no published proceedings. The sixth edition of IWCV will be held in Modena in Canalgrande Hotel, Italy, from the 29th to 31st of May 2018.
Seminar by Univ. of Ferrara - Città educante · 19 Apr 2018
During the final meeting of the partnership between Aimagelab and University of Ferrara within the Città Educante project, Profs Evelina Lamma, Fabrizio Riguzzi and Riccardo Zese hold a seminar about Symbolic Learning an Logic programming. The slides of the seminar are available for downloading (Slide1, Slide2).
Invited talk "L'arte dell'innovazione" · 21 Mar 2018
Sensors - Special Issue on Depth Sensors and 3D Vision · 06 Mar 2018
We are organising a Special Issue entitled "Depth Sensors and 3D Vision" on Sensors and we would like to invite you to contribute one article/review to this special issue. Deadline for manuscript submissions: 31 August 2018.
For further reading, please follow the link to the Special Issue Website at:
Invited talk at the "Technological Innovation for Digital Humanities" workshop · 06 Mar 2018
Lorenzo Baraldi gave an invited talk at the "Technological Innovation for Digital Humanities" workshop, held in Pavia on the 6th of March.
"CultMedia: Deep Learning for automatic description of images and video in DH"
Slides from the talk are available.
Invited talk at AI DIVE 2018 · 27 Feb 2018
Rita Cucchiara gave an invited talk at AI DIVE 2018, which was held in Milan on the 27th of February:
"Computer Vision and Deep Learning: and they lived happily ever after"
Slides from the talk are available.
CVPR 2018 · 20 Feb 2018
"LAMV: Learning to align and match videos with kernelized temporal layers"
by L. Baraldi, M. Douze, R. Cucchiara and H. Jégou
has been accepted to CVPR 2018, that will take place from June 18th to June 22nd in Salt Lake City, Utah.
Why is Deep Learning so cool? · 11 Jan 2018
Il 29 gennaio 2018, presso la Sala Eventi del Tecnopolo di Modena si terrà l'evento:
"Why is Deep Learning so cool?"
Giornata dell'Associazione Italiana per la ricerca in Computer Vision, Pattern Recognition e Machine Learning, GIRPR → VPL
che ospiterà un talk del Prof. Naftali Tishby, Hebrew University of Jerusalem, e l'assemblea straordinaria dell'Associazione.
Ulteriori informazioni sono disponibili alla pagina dedicata all'evento.
Seminar by Dr. Hamed R. Tavakoli · 12 Dec 2017
Dr. Hamed Rezazadegan Tavakoli, Postdoctoral Researcher at Aalto University (Finldand) will give a talk about his research activities on Tuedsay, 12 at 17 pm. The talk will take place in the meeting room of building MO-52.
Dr. Hamed R. Tavakoli has investigated visual attention mechanism and saliency modeling along with their contribution to machine vision algorithms for different applications, including, fixation prediction, object detection and tracking, learning image features from image statistics and egocentric vision. He has also researched inferences from eye movements as a medium that may reveal one’s mind. In hisresearch, machine learning techniques within the span of Bayesian approaches and neural methods, in specific deep learning, have been developed and employed.
Talk at TEDxModenaSalon · 22 Nov 2017
Invited Talk at Intel Nervana AI Academy · 15 Nov 2017
AImageLab will give a talk at the "Deep Learning for scientific research" workshop organized by Intel and Cineca.
Deep learning for Automatic Video Annotation and Captioning
Rita Cucchiara and Lorenzo Baraldi
Automatic Video analysis can reach terrific results in structural and semantic annotation of content, tagging, similarity search, saliency detection and textual captioning. We will describe results achieved in this fileds with specifically designed convolutional neural networks and recurrent networks such as LSTM, trained over very large datasets of broadcast video in sport, movies and culture.
Seminario Prof. Simone Arcagni - · 27 Sep 2017
Venerdì 29 Settembre dalle ore 16 alle ore 17 presso la Sala Riunioni del Tecnopolo (Primo Piano) si terrà un seminario tenuto dal Prof. Simone Arcagni.
L'OCCHIO DELLA MACCHINA
"La scienza non ha nessun bisogno della filosofia per i suoi compiti. In compenso, quando un oggetto è scientificamente costruito con funzioni, per esempio una geometria spaziale, rimane cercare la filosofia che non è assolutamente attribuita alla funzione".
(Deleuze, Guattari, Che cos'è la filosofia?, Einaudi, Torino, 1996)
Lo sguardo e la visione sono questioni che attengono tanto agli studi sui media, alla filosofia e agli studi culturali, quanto a quelli legati alla scienza e alle tecnologie.
Questo seminario sarà un'introduzione alla logica culturale dello sguardo contemporaneo, agli aspetti culturali e alle relazioni sociali dell'occhio computazionale.
Alcune storie di incontri tra studi scientifici e studi umanistici dimostrano come sia fondamentale per entrambi i campi scambiare punti di vista e modelli di analisi per poter affrontare le proprie ricerche con strumentazioni sempre più adatte al soggetto di studi.
Sicurezza Stradale: Imagelab premiato a livello nazionale dalla fondazione ANIA · 22 Sep 2017
La tesi intitolata "Pose estimation tramite tecniche di deep learning per automotive", sviluppata all'interno del laboratorio ImageLab, ha vinto il premio nazionale ANIA come migliore lavoro dell'area tecnico-ingegneristica riguardante il tema della sicurezza stradale. Autore della tesi è Marco Venturelli, relatore la Prof.ssa Rita Cucchiara, correlatori il Prof. Roberto Vezzani e il dott. Guido Borghi.
Tale premio era riservato alle tesi di laurea magistrali e triennali discusse tra il 1° novembre 2015 e il 31 marzo 2017, ed è nato con l’idea di investire e di creare un’interazione con i giovani laureati che abbiano trattato ed approfondito, ognuno nel proprio campo d’interesse, il tema della sicurezza stradale, dando nuovi ed interessanti spunti.
Link alla notizia pubblicata su Repubblica: clicca qui
Clicca qui per maggiori informazioni riguardo al progetto di ricerca.
We are glad to announce that the Master's Thesis "Pose estimation tramite tecniche di deep learning per automotive" developed at ImageLab by Marco Venturelli, Rita Cucchiara, Roberto Vezzani and Guido Borghi has won the national contest about Road Security, promoted by ANIA.
News Link: click here
Click here for more info about the research project.
Rita Cucchiara's invited seminar to Stanford AI lab · 24 Jul 2017
Imagelab ranks first at the LSUN challenge @ CVPR17 · 20 Jul 2017
The latest saliency prediction model developed by Imagelab has ranked first in the LSUN 2017 saliency challenge.
Our model integrates an LSTM-based attentive mechanism to iteratively attend and refine predictions at different locations. A variation of that model is currently under submission to a journal. You can download the preprint of the paper here.
Thanks to: Marcella Cornia, Lorenzo Baraldi, Giuseppe Serra, Rita Cucchiara
MuMeT - Happy Hour Opening Ceremony · 16 Jun 2017
The second edition of a unique Master in Italy, with excellent career perspectives: advanced skills in Visual Computing, Computer Vision and Graphics, Machine Learning and AI. The Master is founded by MIUR and EC in Cluster Smart City and Communities.
Alessandro Capra (Director of Dep. Engineering “Enzo Ferrari”) and Rita Cucchiara (Master Director) will open the event with their greetings. Two welcome speakers Giorgio de Mitri (Sartoria Comunicazione) and Pietro Altoé (NVIDIA) will participate to the happy hour.
Visit event page for more details.
Imagelab @ International Vehicles Symposium (IV 2017), Los Angeles, California · 09 Jun 2017
Unimore Imagelab is attending at the "Intelligent Vehicles Symposium", one of the most important international conference about next smart vehicle generation that will be held at Redondo Beach, Los Angeles, California, from 11 to 14 June 2017.
Two PhD students, Stefano Alletto and Guido Borghi, are presenting two papers:
- Learning Where to Attend Like a Human Driver (A. Palazzi, F. Solera, S. Calderara, S. Alletto, R. Cucchiara)
- Embedded Recurrent Network for Head Pose Estimation in Car (G. Borghi, R. Gasparini, R. Vezzani, R. Cucchiara)
These works have been carried out within the project "FAR2015 - Monitoring the car driver’s attention with multisensory systems, computer vision and machine learning" funded by the University of Modena and Reggio Emilia.
We also acknowledge the CINECA award under the ISCRA initiative, for the availability of high-performance computing resources and support.
For more information please visit:
Unimore in Israele a parlare di trasferimento tecnologico in high tech · 25 May 2017
Bologna, 12 aprile; Prof. Rita Cucchiara Computer VIsion for Industry in The Deep Learning Era · 11 Apr 2017
Two papers accepted @IV 2017 · 16 Mar 2017
- Learning Where to Attend Like a Human Driver (A. Palazzi, F. Solera, S. Calderara, S. Alletto, R. Cucchiara)
- Embedded Recurrent Network for Head Pose Estimation in Car (G. Borghi, R. Gasparini, R. Vezzani, R. Cucchiara)
have been accepted in IEEE Intelligent Vehicles Symposium (IV 2017).
Two papers accepted @CVPR · 09 Mar 2017
- POSEidon: Face-from-Depth for Driver Pose Estimation (G. Borghi, M. Venturelli, R. Vezzani, R. Cucchiara)
- Hierarchical Boundary-Aware Neural Encoder for Video Captioning (L. Baraldi, C. Grana, R. Cucchiara)
have been accepted in CVPR 2017, that will take place at the Hawaii Convention Center from July 21 to July 26, 2017 in Honolulu, Hawaii.
In the meanwhile the e-version will be available, please take a look at the preliminar versions of the papers.
Seminario · 01 Mar 2017
Machine Vision in the deep learning era · 01 Feb 2017
Ingegneria dei Sistemi Intelligenti: Vision, Deep Learning e Cognitive Computing · 25 Jan 2017
Second edition of Master MuMeT on Visual Computing and Multimedia Technologies is starting! · 25 Jan 2017
MuMeT 2017 is an international Master of second level of the University of Modena and Reggio Emilia, co-sponsored by Italian MIUR (Ministero della Istruzione, Università e Ricerca) in the National Technology Clusters (project CittaEducante). MUMET II level Master wants to create new professional figures with advanced knowledge on the emerging technologies in visual computing, vision and machine learning, and intelligent multimedia systems.
The Master homepage can be found here.
Il fattore umano nelle auto a guida autonoma. Attenzione umana ed intelligenza artificiale · 18 Jan 2017
Invited talk "human behavior understanding in automotive and around" · 07 Nov 2016
Invited talk "Visual Intelligence For Human And Machine" · 28 Oct 2016
Multi-Camera Tracking: following people in large camera networks · 14 Oct 2016
Dr. Ergys Ristani is visiting ImageLab and he will give a talk on Multi-Camera Tracking on October, Tuesday 18th at 4pm in FA-0A
Multi-Camera Tracking is a canonical problem in computer vision which tries to determine "who is where" at all time across a large camera network. Typical applications include city-wide surveillance and airport/station monitoring.
Dr. Ergys Ristani is a PhD candidate at Duke University, under the supervision of Prof. Carlo Tomasi. His research interest include Multi-Target Tracking, Multi-Camera Tracking, and the estimation and tracking of motion boundaries.
To download the seminar flyer, please see: https://aimagelab.ing.unimore.it/imagelab/uploadedFiles/seminario_ristani_18_10_2016.pdf.
Spotting prejudice team @Imagelab press release · 03 Oct 2016
The Prejudice team of unimore composed by Imagelab memebers Andrea Palazzi, Simone Calderara and Rita Cucchiara has gained the attention of the scientific community with their recent work about "Spotting prejudice with nonverbal behaviours" presented at UBICOMP 2016.
Check out our "Prejudice dataset" here.
Attached a Press release of all articles that talk about our work on the web and in printed papers:
https://www.newscientist.com/article/mg23130933-200-camera-spots-your-hidden-prejudices-from-your-body-language/ http://www.dailymail.co.uk/sciencetech/article-3813425/Can-computer-tell-RACIST-Algorithm-detect-hidden-prejudice-person-s-body-language.html http://www.repubblica.it/tecnologia/2016/09/29/news/studio_razzismo_ecco_l_algoritmo_italiano_per_smascherare_i_pregiudizi_anche_inconsapevoli-148760996/ http://myinforms.com/en-gb/a/41993012-camera-spots-your-hidden-prejudices-from-your-body-language/ http://anglenews.com/can-a-computer-tell-if-youre-racist-algorithm-can-detect-hidden-prejudice-from-a-persons-body-language/ http://tech.fanpage.it/l-algoritmo-italiano-che-svela-razzismo-e-pregiudizi/ https://www.tomshw.it/l-algoritmo-italiano-per-smascherare-razzismo-e-pregiudizi-80376 https://www.bcrmagazine.it/201633983/razzismo-ecco-lalgoritmo-italiano-per-smascherare-i-pregiudizi-anche-inconsapevoli.php http://23on.com/razzismo-ecco-lalgoritmo-italiano-per-smascherare-i-pregiudizi-anche-inconsapevoli/ http://www.modicanews.com/can-a-computer-tell-if-youre-racist-algorithm-can-detect-hidden-prejudice-from-a-persons-body-language/ http://huntnews.in/p/detail/2821262581067916?uc_param_str=dnfrpfbivesscpgimibtbmntnijblauputoggdnw&pos=1475155800173&channel=technology&chncat=category_english http://ebalsblog.blogspot.it/2016/09/can-computer-tell-if-youre-racist-new.html http://www.looooker.com/archives/34774 http://newsfeeds.pegitboard.com/news/can-a-computer-tell-if-you-re-racist-algorithm-can-detect-hidden-prejudice-from-a-person-s-body-language?uid=24563
New Scientist features ImageLab research about recognition of non-verbal behaviours · 29 Sep 2016
The international science magazine New Scientist, based in UK, talks about the ImageLab research about recognition of non-verbal behaviours in one of its cover articles. The link to the article for the online version of the science magazine can be found here: https://www.newscientist.com/article/mg23130933-200-camera-spots-your-hidden-prejudices-from-your-body-language/.
ImageLab algorithm in OpenCV · 26 Sep 2016
OpenCV accepted our Connected Components Labeling algorithm into their official source code. From the next release, our contribution will be the official one, recognizing its superior performance.
OpenCV (Open Source Computer Vision Library) is an open source computer vision and machine learning software library. OpenCV was built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in the commercial products. The library has more than 2500 optimized algorithms, which includes a comprehensive set of both classic and state-of-the-art computer vision and machine learning algorithms. These algorithms can be used to detect and recognize faces, identify objects, classify human actions in videos, track camera movements, track moving objects, extract 3D models of objects, produce 3D point clouds from stereo cameras, stitch images together to produce a high resolution image of an entire scene, find similar images from an image database, remove red eyes from images taken using flash, follow eye movements, recognize scenery and establish markers to overlay it with augmented reality, etc. OpenCV has more than 47 thousand people of user community and estimated number of downloads exceeding 7 million. The library is used extensively in companies, research groups and by governmental bodies.
Facebook AI Research Partnership · 29 Aug 2016
Facebook has selected Imagelab as one of the 15 world-class research labs in Europe to receive a GPU-based server as part of the Facebook AI Research Partnership. Our application has been selected from a large pool of universities and research institutes working on important and innovative research. Facebook is committed to building strong research partnerships with institutions throughout Europe, and ensuring that Facebook's partners have the necessary hardware and tools to continue tackling some of the most important challenges in the disciplines of artificial intelligence and machine learning. Thanks to Lorenzo Baraldi, Costantino Grana and Simone Calderara!
Read more on the Facebook Research blog.
Report of Dagstuhl Seminar - Eyewear Computing · 19 Jul 2016
Invited talk alla scuola vismac 16 giugno 2016 · 16 Jun 2016
First International Workshop on Egocentric Perception, Interaction and Computing · 24 May 2016
ImageLab is organising the First International Workshop on Egocentric Perception, Interaction and Computing (EPIC@ECCV16) that will be held on the October 8-10, 2016, in conjunction with the European Conference on Computer Vision, Amsterdam, The Netherlands. Its goal is to give an overview of the recent technologies and system solutions, create a forum to exchange ideas and address challenges emerging in this field.
Talk by Prof. Tal Hassner · 17 May 2016
Prof. Tal Hassner (The Open University of Israel) is visiting Imagelab and is giving a talk on
"Faces, deep learning and the pursuit of training data"
Tuesday May 17, 2016 - 02:00 p.m. - Aula P 0.2 ex FA-0A
Abstract: The abilities of machines to detect and recognize faces improved remarkably over the last few years. This progress can at least partially be explained by the sizes of the training sets used to train deep learning models: huge numbers of face images downloaded and manually labeled. It is not clear, however, if the formidable task of collecting and labeling so many images is truly necessary. I will discuss the problems of data collection and describe a number of effective techniques for maximizing deep learning capabilities when collecting additional data is not an option. Importantly, though this talk will focus on face processing related tasks, these techniques can be applied in other image understanding problems where obtaining enough labeled examples for training deep learning systems is hard.
Nuovi sistemi intelligenti - Il Deep Learning per l'impresa del futuro · 16 Apr 2016
Il 16 maggio 2016 presso l'Auditorium del Tecnopolo di Modena si terrà una giornata dedicata al Deep Learning per l'impresa del futuro. Nel corso dell’iniziativa verranno presentate soluzioni, piattaforme software ed infrastrutture utilizzabili per portare il deep learning e le tecniche avanzate di apprendimento automatico nell’impresa, con testimonianze internazionali, accademiche e industriali. Seguirà una tavola rotonda sul loro utilizzo nell'industria e nell'impresa del futuro.
Per informazioni e registrazione: Pagina dedicata all'evento
Startcup regionale e PNI nazionale 2016 · 12 Apr 2016
Il 26 aprile si aprirà il bando per la Startcup Regionale, la competizione per favorire la nascita delle nuove imprese ad alto contenuto innovativo. Si rivolge a persone singole o gruppi interessati a sviluppare idee imprenditoriali innovative e/o ad alto grado di conoscenza. Le migliori idee imprenditoriali parteciperanno al Premio Nazionale Innovazione PNI Italia che si terrà a Modena ospitato da UNIMORE a dicembre 2016.
Per informazioni e contatti: http://www.unimore.it/evidenza/pni2016.html
Talk at "R&D: da Giovani Idee, Grandi Progetti" · 08 Apr 2016
In the context of the talk serie "R&D: da Giovani Idee, Grandi Progetti", Imagelab and Samsung will give a joint talk on 14/04/2016. Ing. Alletto Stefano will present the latest research efforts in the field of wearable computing, internet of things and object recognition. For further information please refer to the talk organizers' website: www.ideelab.it
Rita Cucchiara's interview at TGR TV News · 16 Mar 2016
Prof. Rita Cucchiara explains to the journalists of TV news TGR some of the last results obtained in the project Città Educante, in the context of exploting the new technologies to favour the acceptance of diversity in the primary schools.
The video can be found here.
Rita Cucchiara's talk at Area Chiar CVPR Workshop Vancouver 29 Feb 2015 · 01 Mar 2016
Hardware Grants from Nvidia and Cineca · 10 Feb 2016
Imagelab has received two important hardware grants:
The NVIDIA Hardware Grant, with the donation of one Tesla K40 GPU.
The Italian Supercomputing Resource Allocation (ISCRA) Grant from CINECA, which gives Imagelab access to the Galileo HPC Platform, containg 16 Tesla K80 GPUs.
The granted hardware will be mainly employed in the research project "Deep Learning in videos: concept detection and temporal video segmentation", as well as for other Imagelab activities.
Computer Vision Foundation: Advisory board · 04 Feb 2016
The Computer Vision Foundation appointed the new Advisory Board, chaired by Rene Vidal, composed by Anthony Hoogs, David Forsyth, Dimitri Metaxas,
Josef Sivic ,Kyoung Mu Lee ,Kyros Kutulakos ,Martial Hebert ,Rita Cucchiara ,Ruzena Bajcsy, Stan Sclaroff, Sven Dickinson, and Philip Torr.
Seminario Prof. Antonino Mazzeo · 02 Dec 2015
"Metodi, ambienti IDE, architetture e tecnologie per la progettazione dei sistemi digitali embedded".
Abstract: Il seminario intende focalizzarsi, con riferimento a specifici domini applicativi quali quelli dell’automotive dell’aerospazio e del ferroviario, sulla progettazione dei sistemi digitali embedded ad elevata complessità e assoggettati a vincoli di sicurezza, affidabilità, real-time e time to market e, in particolare, sugli aspetti delle differenti soluzioni oggi consentite dalle correnti tecnologie VLSI inquadrate in vari contesti architetturali dedicati (sistemi dedicati, general purpose e misti realizzati su FPGA, SoC e MPSoC ) e sulle metodologie di sviluppo a supporto.
Sono disponibili i lucidi del seminario al seguente indirizzo
Vision for cultural heritage · 25 Nov 2015
Mercoledì 25 e Giovedì 26 Novembre ImageLab/Softech-ICT ospiterà l'evento "Vision for cultural heritage", nell'ambito del progetto Dicet.
Il programma e il materiale dell'evento sono disponibili al seguente link: Vision for cultural heritage
Lectures by Prof. Nadia Magnenat Thalmann · 25 Sep 2015
Tutorial at CAIP 2015 · 01 Sep 2015
Costantino Grana and Giuseppe Serra gave a tutorial at the International Conference on Computer Analysis of Images and Patterns in Malta on The Bag of Visual Words model and recent advancements in image classification. Tutorial slides
International Workshop on Wearable and Ego-vision Systems for Augmented Experience (WEsAX) · 03 Jul 2015
ImageLab is organising the first International Workshop on Wearable and Ego-vision Systems for Augmented Experience (WEsAX) that will be held on the July 3, 2015, in conjunction with the IEEE International Conference on Multimedia and Expo (ICME), Turin, Italy. The goal of the first International Workshop on Wearable and Ego-vision Systems for Augmented Experience (WEsAX) is to give an overview of the recent technologies and system solutions, create a forum to exchange ideas and address challenges emerging in this field.
Lecture by Ing. Alessio Bazzica · 22 May 2015
Ing Alessio Bazzica is visiting Imagelab and is going to present a lecture on
Discovering the Music Information Retrieval Field
(Friday May 22, 2015 - 9:00am - Aula FA-1D )
Alessio Bazzica received the M.S. degree in Computer Engineering from the University of Florence (Italy) in 2012 and is now a 3nd year PhD candidate at the Multimedia Computing (MMC) Group in Delft (The Netherlands). He is interested in multimedia information retrieval and he is currently working on multimodal and crossmodal approaches for music information retrieval.
Attivazione master di II livello - MUMET · 08 May 2015
Al via il nuovo Master MUMET. Il master vuole creare nuove figure professionali (MULTIMEDIA DATA SCIENTIST), con competenze avanzate sulle tecnologie informatiche multimediali e sul visualcomputing, competenze attualmente molto rare ed assai richieste per i sistemi, servizi ed applicazioni nelle smartcities and communities.
Workshop tematico - "New Ideas for Education in Smart Communities" · 20 Apr 2015
Il progetto Città Educante ha lo scopo di ripensare radicalmente l'ambiente di apprendimento, attraverso l'applicazione delle più avanzate tecnologie informatiche. Il workshop, ad invito e a partecipazione libera, ha come obiettivo lo scambio di idee e la disseminazione delle attività svolte dai partners durante il primo anno di progetto.
L'evento che si inserisce tra la manifestazioni per i 25 anni di Ingegneria a Modena si svolgerà lunedì 20 aprile 2015, ore 10:00, presso l'aula P 0.5 (FA-0-D), del Dipartimento di Ingegneria "Enzo Ferrari".
Best Paper award @ AVSS2015 · 08 Jan 2015
Group Detection Paper accepted @TPAMI · 08 Jan 2015
Special Issue on: Wearable and Ego-vision Systems for Augmented Experience · 08 Jan 2015
The rapid progress in the development of low-level component technologies such as wearable cameras, wearable sensors, wearable displays and wearable computers is making it possible to augment everyday living. Wearable and egocentric vision systems can be exploited to analyze multi-modal data types (e.g. video, audio, motion) and to support understanding human interactions with the world (including gesture recognition, action recognition, social interaction recognition). Based on the processing of such data, wearable systems can be used to enhance our capabilities and augment our perception. State-of-the-art techniques for wearable sensing can support assistive technologies and advanced perception. This special issue intends to highlight research in support for human performance through egocentric sensing .
Best paper at INTETAIN 2015 · 08 Jan 2015
The paper “ Wearable Vision for Retrieving Architectural Details in Augmented Tourist Experiences ” by Stefano Alletto, Davide Abati, Giuseppe Serra and Rita Cucchiara was awarded the best paper award at INTETAIN in Turin. In this paper we propose an egocentric vision system to enhance tourists’ cultural heritage experience. Exploiting a wearable board and a glass-mounted camera, the visitor can retrieve architectural details of the historical building he is observing and receive related multimedia contents. To obtain an effective retrieval procedure we propose a visual descriptor based on the covariance of local features. Differently than the common Bag of Words approaches our feature vector does not rely on a generated visual vocabulary, removing the dependence from a specific dataset and obtaining a reduction of the computational cost. 3D modeling is used to achieve a precise visitor’s localization that allows browsing visible relevant details that the user may otherwise miss. Experimental results conducted on a publicly available cultural heritage dataset show that the proposed feature descriptor outperforms Bag of Words techniques.
Seminario Nvidia · 17 Nov 2014
Lunedì 17 Novembre 2014 dalle 14.00 alle 16.15 presso l'Aula P0.4 (ex FA-0C) - dipartimento DIEF - si terrà un seminario sulle architetture parallele GPU Nvidia.
14:00-15:00: Edmondo Orlotti, Business Development NVIDIA "Tecnologie per il calcolo parallelo e la visualizzazione, dal cloud al mobile"
15:00-16:15: Carlo Nardone, Solution Architect NVIDIA, "Gli strumenti di sviluppo per la programmazione in ambiente GPU". Locandina
Slides: Parte 1, Parte 2, Parte 3.
Eighth edition of the ACM/IEEE International Conference on Distributed Smart Cameras (ICDSC 2014) · 04 Nov 2014
Invited talk at the ECCV workshop · 08 Jan 2014
Prof. Cucchiara is presenting a talk on "Computer Vision for interactive experiences with art and artistic documents" at the ECCV workshop "Where Computer Vision Meets Art - VISART2014.
2nd International Workshop on Multimedia for Cultural Heritage · 09 Sep 2013
The 2nd International Workshop on Multimedia for Cultural Heritage, will be held on the 9-10th of September 2013, in conjunction with the 17th International Conference on Image Analysis and Processing (ICIAP), Naples, Italy, aims to be a profitable informal working day to discuss together hot topics in multimedia applied to cultural heritage.
Public website: http://imagelab.ing.unimore.it/MM4CH2013/
S5 - Second Short Spring School in Surveillance (Modena, May 7-9 2013) · 07 May 2013
This short 3-day school is meant to provide a self-contained comprehensive introduction to modern video surveillance methods and techniques, with a good balance between theory and practical applications.
This school is open to researchers, PhD and undergraduate students, scholars in the field of surveillance and security, and is conceived also for technicians from both industries and public entities.
Public website: http://imagelab.ing.unimore.it/s5/
3dPes · 12 Jan 2013
3DPeS (3D People Surveillance Dataset) is a new surveillance dataset, designed mainly for people re-identification in multi camera systems with non-overlapped field of views, but also applicable to many other tasks, such as people detection, tracking, action analysis and trajectory analysis.
Differently from other re-identification datasets here data for the complete processing chain are available: the camera setting and the 3D environment reconstruction, the hundreds of recorded videos, the camera calibration parameters, the identity of the hundreds of people, detected more than one time by different point of view.
Dataset web site: 3DPeS
S4 - Short Spring School in Surveillance · 17 May 2011
This short 3-day school is meant to provide a self-contained comprehensive introduction to modern video surveillance methods and techniques, with a good balance between theory and practical applications.
This school is open to PhD and undergraduate students, scholars in the field of surveillance and security, and is conceived also for technicians from both industries and public entities.
Public website: http://imagelab.ing.unimore.it/s4/
Multimedia in Forensics - MiFor2009 · 23 Oct 2009
With the proliferation of multimedia data on the web, surveillance cameras in cities, and mobile phones in everyday life we see an enormous growth in multimedia data that needs to be analyzed by forensic investigators. The sheer volume of such datasets makes manual inspection of all data impossible. Tools are needed to support the investigator in their quest for relevant clues and evidence and in their strive towards preventing crime.
The multimedia community has developed new solutions for management of large collections of video footage, images, audio and other multimedia content, knowledge extraction and categorization, pattern recognition, indexing and retrieval, searching, browsing and visualization, and modeling and simulation in various domains. Due to the inherent uncertainty and complexity of forensic data, applying those techniques to forensic data is not straightforward.
The time is ripe to tailor these results for forensics. Multimedia in forensics is the workshop aims at joining the topics and their communities.
This workshop aims at addressing the multimedia toolbox supporting the forensic process from the prevention of crime, capturing and annotation of the crime scene, the investigation of the data in the lab, up to the presentation of the results in court. It is a first attempt in bringing multimedia tools in to this exciting application field. Target audience are researchers working on innovative technology, representatives from companies developing tools, and forensic investigators in various disciplines
Presente e Futuro dei sistemi di Videosorveglianza per la sicurezza urbana · 20 Feb 2009
Il Comune di Modena e il Dipartimento di Ingegneria dell'Informazione dell'Università di Modena e Reggio Emilia organizzano il convegno "Presente e Futuro della Videosorveglianza per la sicurezza urbana" il giorno 20 Febbraio 2009 presso la sala della Fondazione Biagi.
L'argomento è la videosorveglianza e il trattamento dei dati video sia in tempo reale, per la prevenzione e la salvaguardia del cittadini, sia a posteriori per l'analisi forense e il supporto nelle fasi processuali.
Iciap 2007 · 12 Sep 2007
VSSN 2006 · 27 Oct 2006
Following a successful tradition, this will be the fourth edition of the VSSN Workshop to be held in conjunction with ACM Conference of Multimedia 2006 at Santa Barbara, California, October 22-28, 2006.
Workshop web site: http://imagelab.ing.unimore.it/vssn06
VSSN 2005 · 11 Nov 2005
Rita Cucchiara was program chair of VSSN 2005, the 3rd ACM International Workshop on Video Surveillance & Sensor Networks in conjunction with ACM Multimedia 2005.
Event web site: http://imagelab.ing.unimore.it/vssn05/