Research projects

Multimedia

  • H2020 AI4Media - Senior researcher, General Assembly member - "A European Excellence Centre for Media, Society and Democracy", H2020 ICT-48-2020 / Towards a vibrant European network of AI excellence centres, 2020-2024

Artificial data generation:
  • Keysight Romania - Researcher - "Machine Learning Techniques for Generating Network Traffic Data", 2020. Industry project

Surveillance and security:
  • PIMS-IAT - Researcher - "Platform Specialized in Identifying and Evaluating Early Warning Indices for Crisis Management", Solutions Axis, grant 27SOL/2021, 2021-2023
  • DeepVisionRomania - Researcher - "Identifying People in Video Streams using Silhouette Biometrics", Solutions Axis, grant 28SOL/2021, 2021-2023
  • IDENEO - Researcher - "Identifying the Unusual Behavior of People in Video Streams", Solutions Axis, grant 29SOL/2021, 2021-2023
  • GRAVI - Researcher - "Virtual Guardian: Artificial Intelligence Powered Multi-Sensor System for Automatic Securing of Areas of Interest", Industry Transfer Axis, grant PN-III-P2-2.1-PTE-2019-0570, 2020-2021
  • SPIA-VA - Researcher - "Technologies and Innovative Video Systems for Person Re-Identification and Analysis of Dissimulated Behavior", Solutions Axis, grant 2SOL/2017, 2017-2020
  • SPOTTER - Researcher - "Real-time IP Camera-based Intelligent Video Surveillance Security System with DROP Retrieval", PED Axis, grant 30PED/2017, 2017-2018
  • SCOUTER - Researcher - "Smart Video Surveillance System with Automatic Localization of Target Events", Innovation Axis, grant 28DPST/2013, 2013-2015

Hardware-Software (Co-)Design and Embedded deep learning:
  • Keysight Romania - Researcher - "AI for Hardware-Software Co-Design", 2024. Industry project
  • NXP Semiconductors Romania - Researcher - "RISC V-based Hardware-Software System for Machine Learning Applications", 2019, Industry project

Retail services
  • SMARTRetail - Researcher - "Enhancing and Improving Customer Experience and Services in Supermarkets via SMART Artificial Intelligence Powered Systems", Industry Transfer Axis, grant PN-III-P2-2.1-PTE-2019-0055, 2020-2021

Cultural herritage:
  • UMETECH - Researcher, Trainer - "University & Media Technology for Cultural Heritage", 2017-2019

Spectroscopy analysis:
  • THz-DETECT - Researcher - "Millimeter Wave (THz) Time Domain Spectroscopy Instrument for Dangerous Substances Identification", PN-II-PT-PCCA-2013-4-0967, 2014-2017

Datasets
  • Interestingness10K - The dataset is designed for the task of predicting multimedia interestingness in images and videos, corresponding with the different editions of the MediaEval Predicting Media Interestingness task. Data consists of movie excerpts and key-frames and their corresponding ground-truth files based on the classification into interesting and non-interesting samples, along with a set of pre-processed descriptors. A thourough analysis of this dataset, MediaEval participant methods and feature performance analysis, performance enhancement suggestions and much more can be found in this IJCV paper.
  • VSD96 - The dataset is designed for the task of detecting violent events in movies and web videos, and is delivered in several packages, corresponding with the different editions of the MediaEval Affect Task - Violent Scenes Detection benchmarking campaign. Each video comes with ground-truth annotations, pre-extracted visual features and concept annotations. A thorough analysis of this dataset, MediaEval participant methods and features performance analysis and much more can be found in this TAffC paper.
  • Predicting Media Memorability 2019 - The dataset is designed for the task of predicting media memorability, and it was validated during the 2019 Predicting Media Memorability Tasks at the MediaEval Benchmarking Initiative for Multimedia Evaluation. It consists of 10000 short videos, split between a 8000-video devset and a 2000-video testset. Each video comes with a set of visual features and titles than can be interpreted as textual features.
  • MMTF-14K - The MMTF-14K dataset provides a stable and extensive source for devising and evaluating movie recommender systems. MMTF-14K contains audio and visual descriptors in addition to ratings and metadata for 13,623 Hollywood-type movie trailers. The dataset therefore facilitates research on content-based recommender systems, where content refers not only to metadata, but specifically to visual and auditory characteristics of movies. The data comes also with several baselines benchmarking results for uni-modal and multi-modal recommendation systems.
Organizational activities

Editorial board:


Special Sessions and Special Issues organizer
Main organizer for the following tasks:
Part of the organising committee for the following conferences:
  • MediaEval 2023 MediaEval Benchmarking Initiative for Multimedia Evaluation, Amsterdam, The Netherlands, Feb. 2023
  • MediaEval 2022 MediaEval Benchmarking Initiative for Multimedia Evaluation, Bergen, Norway, Jan. 2023
  • MediaEval 2021 MediaEval Benchmarking Initiative for Multimedia Evaluation, Online Event, Dec. 2021
  • CLEF 2021 , Conference and Labs of the Evaluation Forum, Online Event, Sep. 2021
  • MediaEval 2020 MediaEval Benchmarking Initiative for Multimedia Evaluation, Online Event, Dec. 2020
  • MediaEval 2019 MediaEval Benchmarking Initiative for Multimedia Evaluation, Sophia Antipolis, France, Oct. 2019
  • MediaEval 2018 MediaEval Benchmarking Initiative for Multimedia Evaluation, Sophia Antipolis, France, Oct. 2018
  • ICMR 2017 - ACM International Conference on Multimedia Retrieval, Bucharest, Romania, Jun. 2017.
  • MediaEval 2017 MediaEval Benchmarking Initiative for Multimedia Evaluation, Dublin, Ireland, Sep. 2017
  • CBMI 2016 - 14th IEEE/ACM International Workshop on Content-Based Multimedia Indexing, Bucharest, Romania, Jun. 2016.
  • MediaEval 2016 MediaEval Benchmarking Initiative for Multimedia Evaluation, Hilversum, Netherlands, Oct. 2016

Part of the organising comitee for the following tasks

Reviewing activities

Reviewer for the following journals:


TPC, Reviewer or subreviewer for the following conferences and workshops:
  • MAD @ ICMR 2022 - 1st ACM International Workshop on Multimedia AI against Disinformation, organized with ICMR 2022, 27-30 June, Newark, USA.
  • ICMR 2022 - ACM International Conference on Multimedia Retrieval, 27-30 June, Newark, USA.
  • ACMMM 2022 - 30th ACM International Conference on Multimedia, 10-14 October 2022, Lisbon, Portugal.
  • ACMMM 2021 - 29th ACM International Conference on Multimedia, 12-16 October 2021, Chengdu, China.
  • RISS 2020 - 3rd International Workshop on Research & Innovation for Secure Societies, @ICPR 2020, Milan, Italy, Jan. 2021
  • ACMMM 2020 - 28th ACM International Conference on Multimedia, 12-16 October 2020 Seattle, United States
  • ECIR 2020 - The 42nd European Conference on Information Retrieval, Lisbon, Portugal, Apr. 2020
  • ACMMM 2019 Engagement – Multimedia search and recommendation area - 27th ACM International Conference on Multimedia, Nice, France
  • ISM 2019 - The 21st IEEE International Symposium on Multimedia, San Diego, USA, Dec. 2019
  • RISS 2019 - 2nd International Workshop on Research & Innovation for Secure Societies, organized with SpeD 2019, Timisoara, Romania, Oct. 2019
  • MMM 2019 - 25th International Conference on MultiMedia Modeling, Thessaloniki, Greece, Jan. 2019
  • ICMR 2018 Special Session Predicting User Perceptions of Multimedia Content - ACM International Conference on Multimedia Retrieval, ICMR 2018, Yokohama, Japan
  • ISM 2018 - The 20th IEEE International Symposium on Multimedia, Taichung, Taiwan, Dec. 2018

Talks

  • AI Cafe 2022 - AI Cafe: Deep Ensembles for the Prediction of Media Interestingness, May 2022