Home

Research Statement

I am currently (since November 2020) a Tenure-Track Assistant Professor (Italian RTD/B) at the Information Engineering and Computer Science and Mathematics (DISIM) Department of the University of L’Aquila, Italy. Previously, I was an Assistant Professor (Italian RTD/A) at the Faculty of Computer Science of the Free University of Bozen-Bolzano, Italy. I received the master's and PhD degrees in Computer Science and Engineering from the University of Naples “Federico II”, Italy, in 2009 and 2013, respectively, under the supervision of Professors Antonio Picariello and Vincenzo Moscato. During my PhD, I was visiting scholar at the University of Maryland, College Park (USA) for six months, under the supervision of Prof. V.S. Subrahmanian. I was also Postdoctoral Research Fellow at the Department of Electrical Engineering and Information Technology of the University of Naples "Federico II" from 2013 to 2015. Moreover, I achieved the Italian National Scientific Habilitation (ASN) as Associate Professor in the scientific-disciplinary sectors ING-INF/05 and INF/01 in November 2020.

My research interests include the fields of event detection and analysis in large scale databases applied to multimedia, semantic analysis and security, as demonstrated by several papers published in top journals (such as TKDE, TOIT, and VLDBJ) and conferences (IJCAI, EDBT, CIKM). During my PhD, I mostly worked on the detection of unexplained activities in a sequence of time-stamped observation data [1, 2], under the supervision of Professors Antonio Picariello, Vincenzo Moscato and V.S. Subrahmanian; more specifically, we defined a general model along with some complex algorithms to detect unexplained activities, and we carried out massive experiments in both video surveillance and cyber-security domains. Afterwards, we have also extended such approach to social network analysis [4].

Since I arrived to Bolzano (May 2015), I worked on a highly interactive event detection system in the video surveillance context in collaboration with Prof. Sven Helmer. Such system ranges from the generation of raw events to the formulation and the detection of complex high-level events in both offline and online modes. We presented the overall system at CIKM 2017 in Singapore [5, demo]. More specifically, the system is based on an extension of relational algebra, ISEQL (Interval-based Surveillance Event Query Language) - that we defined in a specific paper [6] -, enriched with powerful temporal operators. Such operators have been implemented by defining some very efficient plane sweeping interval join algorithms [3]. Moreover, we have recently also adapted and extended ISEQL to fit the requirements of a social network setting. In fact, we have extended the overall event detection framework to allow the detection of high-level events in a social network context [10], for instance abnormal or malicious behavior such as spamming. Additionally, I also worked on a probabilistic approach for itinerary planning with category constraints (in collaboration with Dr. Paolo Bolzoni and Prof. Sven Helmer) [7], and on a system including social sensing for improving the user experience in orienteering [8, 9]. Additionally, I was able to attract more than 45,000 euros in funding to finance part of my research work.

One of the main goals of my research will deal with the integration of the new ISEQL operators defined for social network context in the query processing framework of the PostgreSQL database management system. This involves both further extending the ISEQL interval relationships – in order to make them even more flexible - , and improving their efficiency – in order to further speed-up the detection process, especially in a real time setting. Additionally, I also plan to integrate the overall framework for high-level surveillance event detection with a smart robot, in order to enhance the monitoring process. More specifically, the robot will be able to improve the quality of a video stream, which in the earlier work was recorded from a static camera position, by providing a mobile camera, enabling us to get footage from different angles. Furthermore, the robot platform will be able to go beyond just contacting security or the police by taking immediate action when a potentially dangerous event is detected: for instance, picking up an unattended package. Preliminary versions of this framework prototype are shown in [11] and [12]. Then, the following goal would be to improve and extend our current work on event-detection platforms. More specifically, I plan to further increase the efficiency of the data processing components, allowing for faster reaction times and a higher data throughput. Additionally, I want to improve the video capturing capabilities by switching to a multi-camera configuration. Eventually, I plan to replace the Rapiro toy robot that was used for the feasibility studies with more sophisticated hardware to enhance the mobile capabilities of the platform.

Selected Publications

[1] Massimiliano Albanese, Cristian Molinaro, Fabio Persia, Antonio Picariello, VS Subrahmanian: Finding “unexplained” activities in video. Twenty-Second International Joint Conference on Artificial Intelligence (IJCAI'11), Barcelona, July 2011: 1628-1634.

[2] Massimiliano Albanese, Cristian Molinaro, Fabio Persia, Antonio Picariello, VS Subrahmanian: Discovering the Top-k Unexplained Sequences in Time-Stamped Observation Data. IEEE Transactions on Knowledge and Data Engineering, Volume 26, Number 3, 2014: 577-594.

[3] Danila Piatov, Sven Helmer, Anton Dignös, Fabio Persia, Cache-Efficient Sweeping-Based Interval Joins for Extended Allen Relation Predicates, accepted for publication in the VLDB Journal.

[4] Flora Amato, Aniello Castiglione, Aniello De Santo, Vincenzo Moscato, Antonio Picariello, Fabio Persia, Giancarlo Sperlí: Recognizing human behaviours in online social networks. Computers & Security, Volume 74, 2018: 355-370.

[5] Fabio Persia, Fabio Bettini, Sven Helmer: An Interactive Framework for Video Surveillance Event Detection and Modeling. Proceedings of the 2017 ACM on Conference on Information and Knowledge Management (CIKM’17), Singapore, November 2017: 2515-2518.

[6] Sven Helmer, Fabio Persia: ISEQL: an Interval-based Surveillance Event Query Language. International Journal of Multimedia Data Engineering and Management (IJMDEM), Volume 7, Number 4, 2016: 1-21.

[7] Paolo Bolzoni, Fabio Persia, Sven Helmer: Itinerary Planning with Category Constraints using a Probabilistic Approach. International Conference on Database and Expert Systems Applications (DEXA’17), Lyon, August 2017: 363-377.

[8] Fabio Persia, Sven Helmer, Sergejs Pugacs, Giovanni Pilato: Social Sensing for Improving the User Experience in Orienteering, 2019 IEEE 13th International Conference on Semantic Computing (ICSC ‘19), Newport Beach, California, January 2019: 239-246.

[9] Fabio Persia, Giovanni Pilato, Mouzhi Ge, Paolo Bolzoni, Daniela D’Auria, Sven Helmer, Improving Orienteering-based Tourist Trip Planning with Social Sensing, “Data Exploration in the Web 3.0 Age” on the Future Generation Computer Systems (FGCS) Journal 110, 931-945 (2020).

[10] Fabio Persia, Sven Helmer: A Framework for High-Level Event Detection in a Social Network Context Via an Extension of ISEQL, in: IEEE, 2018 IEEE Twelfth International Conference on Semantic Computing (IEEE ICSC 2018),  140-147, (2018).

[11] Daniela D'Auria, Fabio Persia, Fabio Bettini, Sven Helmer, Bruno Siciliano: SARRI: a SmArt Rapiro Robot Integrating a framework for automatic high-level surveillance event detection, in: 2018 Second IEEE International Conference on Robotic Computing (IEEE IRC 2018), 238-241, (2018).

[12] Daniela D'Auria, Fabio Persia, Fabio Bettini, Sven Helmer: Predicting and Preventing Dangerous Events via Video Surveillance Using a Robotic Platform, 2019 Third IEEE International Conference on Robotic Computing (IEEE IRC 2019), 549-554, (2019).

 

Non ci sono annunci disponibili.

Utilizziamo i cookie per offrirti il ​​nostro servizio. Continuando a utilizzare questo sito acconsenti al nostro utilizzo dei cookie come descritto nella nostra policy.