COntactlesS Multibiometric mObile System (in the wild)
The ever-growing number of travelers and migrants crossing the EU borders poses a serious challenge to the border control authorities in terms of a reduced amount of time for carrying out border checks. Consequently, efforts are being undertaken to facilitate the travel of bona-fide and genuine passengers and, at the same time, to safeguard a high level of security. In this kind of context, the use of multimodal biometrics might provide the key for increasing the level of security while reducing the failures inevitably associated with the use of a single identifier in a typically uncontrolled environment. As a further element to consider, practical experiences lead to privilege a most fluent and non-intrusive control process for non-critical travelers (EU, bona-fide etc.). Therefore the use of contactless capturing techniques and their implementation on consumer-level mobile devices is likely to be preferred over contact-based technologies and dedicated devices. Overall, COSMOS (COntactlesS Multibiometric mObile System in the wild) aims at delivering a comprehensive approach to multi-biometric person verification and recognition, including most contact-less biometrics, flexibly integrated through a context-adaptive acquisition/matching strategy based on their complementarity and exploiting the agile and ubiquitous hardware platforms represented by last generation smartphones and tablets. More in practice, the project will exploit the specific knowledge of each of the participants to provide an unprecedented unified biometric platform for contactless person verification/recognition by means of both hard biometrics like face (both in 2D and 3D), iris, ear, fingerprint/palmprint and soft biometrics like gait and gaze. Moreover, multi-tracking methods will be also developed to enabling screening-from-distance capabilities to allow the proposed system to detect subjects of interest or potential threats to be checked in detail by the other biometric modalities. COSMOS is expected to fostering the research and application of new ideas in the field of biometry by providing three major contributions to the field: the effective and efficient implementation of the single modalities on mobile architectures in the challenging "in-the-wild" scenario; the complimentary integration of these biometrics by innovative data fusion strategies to maximize the discriminanting potential of the different identifiers considered according to a wide range of operative conditions and the novel smart management of the crucial privacy issues related to a multi-biometric system. Finally, since in Italy as well in other EU and extra-EU countries, strict data protection prescriptions regulate the use of biometrics, COSMOS will devote a specific emphasis to data protection, social, medical and ethical issues.
The UNIMORE RU will be in charge of developing a flexible MTT solution working with mobile cameras. In particular the problem of tracking multiple target will be studied in an unconstrained scenario with a limited set of computational resources (e.g. processing on a tablet device) and a possibly different set of input cameras (e.g. tablet camera, GoPro Cameras, video streams from UaV). The overall approach consists in creating a joint framework that performs detection and tracking simultaneously, according to the features of the recorded scene. This new model will split the person in parts and track them simultaneously; missing/occluded parts are estimated on the location of observed parts maximizing the tracking confidence score. The objective is to dispose frame by frame of at least one possible target’s part exploitable for biometric recognition.
|1||Fabbri, Matteo; Lanzi, Fabio; Calderara, Simone; Palazzi, Andrea; Vezzani, Roberto; Cucchiara, Rita "Learning to Detect and Track Visible and Occluded Body Joints in a Virtual World" Proceedings of the 15th European Conference on Computer Vision (ECCV) 2018, Munich (Germany), September, 8-14 2018, 2018 Conference|
|2||Fabbri, Matteo; Calderara, Simone; Cucchiara, Rita "Generative Adversarial Models for People Attribute Recognition in Surveillance" Proceedings of the 14th IEEE International Conference on Advanced Video and Signal based Surveillance, Lecce, Italy, 29th August 1st September, 2017, 2017 Conference|