Kernelized Covariance for Action Recognition

J. Cavazza, A. Zunino, M. San Biagio and V. Murino: "Kernelized Covariance for Action Recognition". ICPR 2016 PDF


The descriptive power of the covariance matrix is limited in capturing linear mutual dependencies between variables only. To solve this issue, we present a rigorous and principled mathematical pipeline to recover the kernel trick for computing the covariance matrix, enhancing it to model more complex, non-linear relationships conveyed by the raw data. Our proposed encoding (Kernelized-COV) generalizes the original covariance representation without compromising the efficiency of the computation. Despite its broad generability, the aforementioned paper applied Kernelized-COV to 3D action recognition from MoCap data.


A Game-Theoretic Probabilistic Approach for Detecting Conversational Groups

S. Vascon, E. Zemene, M. Cristani, H. Hung, M. Pelillo and V. Murino: "A Game-Theoretic Probabilistic Approach for Detecting Conversational Groups". ACCV 2014 PDF

A standing conversational group (also known as F-formation) occurs when two or more people sustain a social interaction, such as chatting at a cocktail party. Detecting such interactions in images or videos is of fundamental importance in many contexts, like surveillance, social signal processing, social robotics or activity classification. This paper presents an approach to this problem which models the socio-psychological concept of an F-formation. Essentially, an F-formation defines some constraints on how subjects have to be mutually located and oriented. We develop a game-theoretic framework, embedding these constraints, which is supported by a statistical modeling of the uncertainty associated with the position and orientation of people. Specifically, we propose two novel ways of handling the uncertainty in the position and orientation of the head of each individual with respect to the remaining members of the group in terms of tracking errors and the true body orientation. First, we use a novel representation of the affinity between candidate pairs by expressing the distance between distributions over the most plausible oriented region of attention. Additionally, we integrate temporal information over multiple frames while taking into account the social context established in previous frames. We do this in a principled way by using recent notions from multi-payoff evolutionary game theory. Experiments on several benchmark datasets consistently show the superiority of the proposed approach over state of the art, and its robustness under severe noise conditions.

Multi-Link Analysis: Case-Control Brain Network Comparison via Sparse Connectivity Analysis

A. Crimi, L. Giancardo, F. Sanbataro, A. Gozzi, V. Murino, D. Sona


c fig0

Imaging neuroscience is currently pushing towards the analysis of the brain from a connectivity perspective, which is unveiling many insights into brain structure and functionality. Neuroscientists are often required to evaluate experimental effects in case-control on specimens with thousands connections. This software is based on an unsupervised machine-learning algorithm, which can capture the multivariate relationships that characterize two distinct groups of connectomes, thus allowing neuroscientists to immediately visualize only the sub-networks that contain information about differences between case and control group. The method exploits recent machine learning techniques which employ sparsity in order to deal with weighted network composed of hundreds of thousands of connections.


Heterogeneous Auto-Similarities of Characteristics (HASC): Exploiting Relational Information for Classification

M. San Biagio, M. Crocco, M. Cristani, S. Martelli and V. Murino: "Heterogeneous Auto-Similarities of Characteristics (HASC): Exploiting Relational Information for Classification". ICCV 2013 PDF

Capturing the essential characteristics of visual objects by considering how their features are inter-related is a recent philosophy of object classification. We embed this principle in a novel image descriptor, dubbed Heterogeneous Auto-Similarities of Characteristics (HASC). HASC is applied to heterogeneous dense features maps, encoding linear relations by covariances and nonlinear associations through information-theoretic measures such as mutual information and entropy. In this way, highly complex structural information can be expressed in a compact, scale invariant and robust manner.



IIT's website uses the following types of cookies: browsing/session, analytics, functional and third party cookies. Users can choose whether or not to accept the use of cookies and access the website.
By clicking on further information, the full information notice on the types of cookies used will be displayed and you will be able to choose whether or not to accept cookies whilst browsing on the website.
Further information
Accept and close

I numeri di IIT

L’Istituto Italiano di Tecnologia (IIT) è una fondazione di diritto privato - cfr. determinazione Corte dei Conti 23/2015 “IIT è una fondazione da inquadrare fra gli organismi di diritto pubblico con la scelta di un modello di organizzazione di diritto privato per rispondere all’esigenza di assicurare procedure più snelle nella selezione non solo nell’ambito nazionale dei collaboratori, scienziati e ricercatori ”.

IIT è sotto la vigilanza del Ministero dell'Istruzione, dell'Università e della Ricerca e del Ministero dell'Economia e delle Finanze ed è stato istituito con la Legge 326/2003. La Fondazione ha l'obiettivo di promuovere l'eccellenza nella ricerca di base e in quella applicata e di favorire lo sviluppo del sistema economico nazionale. La costruzione dei laboratori iniziata nel 2006 si è conclusa nel 2009.

Lo staff complessivo di IIT conta circa 1440 persone. L’area scientifica è rappresentata da circa l’85% del personale. Il 45% dei ricercatori proviene dall’estero: di questi, il 29% è costituito da stranieri provenienti da oltre 50 Paesi e il 16% da italiani rientrati. Oggi il personale scientifico è composto da circa 60 principal investigators, circa 110 ricercatori e tecnologi di staff, circa 350 post doc, circa 500 studenti di dottorato e borsisti, circa 130 tecnici. Oltre 330 posti su 1400 creati su fondi esterni. Età media 34 anni. 41% donne / 59 % uomini.

Nel 2015 IIT ha ricevuto finanziamenti pubblici per circa 96 milioni di euro (80% del budget), conseguendo fondi esterni per 22 milioni di euro (20% budget) provenienti da 18 progetti europei17 finanziamenti da istituzioni nazionali e internazionali, circa 60 progetti industriali

La produzione di IIT ad oggi vanta circa 6990 pubblicazioni, oltre 130 finanziamenti Europei e 11 ERC, più di 350 domande di brevetto attive, oltre 12 start up costituite e altrettante in fase di lancio. Dal 2009 l’attività scientifica è stata ulteriormente rafforzata con la creazione di dieci centri di ricerca nel territorio nazionale (a Torino, Milano, Trento, Parma, Roma, Pisa, Napoli, Lecce, Ferrara) e internazionale (MIT ed Harvard negli USA) che, unitamente al Laboratorio Centrale di Genova, sviluppano i programmi di ricerca del piano scientifico 2015-2017.

IIT: the numbers

Istituto Italiano di Tecnologia (IIT) is a public research institute that adopts the organizational model of a private law foundation. IIT is overseen by Ministero dell'Istruzione, dell'Università e della Ricerca and Ministero dell'Economia e delle Finanze (the Italian Ministries of Education, Economy and Finance).  The Institute was set up according to Italian law 326/2003 with the objective of promoting excellence in basic and applied research andfostering Italy’s economic development. Construction of the Laboratories started in 2006 and finished in 2009.

IIT has an overall staff of about 1,440 people. The scientific staff covers about 85% of the total. Out of 45% of researchers coming from abroad 29% are foreigners coming from more than 50 countries and 16% are returned Italians. The scientific staff currently consists of approximately 60 Principal Investigators110 researchers and technologists350 post-docs and 500 PhD students and grant holders and 130 technicians. External funding has allowed the creation of more than 330 positions . The average age is 34 and the gender balance proportion  is 41% female against 59% male.

In 2015 IIT received 96 million euros in public funding (accounting for 80% of its budget) and obtained 22 million euros in external funding (accounting for 20% of its budget). External funding comes from 18 European Projects, other 17 national and international competitive projects and approximately 60 industrial projects.

So far IIT accounts for: about 6990 publications, more than 130 European grants and 11 ERC grants, more than 350 patents or patent applications12 up start-ups and as many  which are about to be launched. The Institute’s scientific activity has been further strengthened since 2009 with the establishment of 11 research nodes throughout Italy (Torino, Milano, Trento, Parma, Roma, Pisa, Napoli, Lecce, Ferrara) and abroad (MIT and Harvard University, USA), which, along with the Genoa-based Central Lab, implement the research programs included in the 2015-2017 Strategic Plan.