Nicoletta Noceti

• Multi-class image classification: sparsity does it better, Sean Ryan Fanello, Nicoletta Noceti, Giorgio Metta, Francesca Odone, VISAPP: International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications., 2013
• Precise people counting in real time, Luca Zini, Nicoletta Noceti, Francesca Odone, IEEE International Conference on Image Processing, 2013
• Background modeling through dictionary learning, Alessandra Staglianò, Nicoletta Noceti, Alessandro Verri, Francesca Odone, IEEE International Conference on Image Processing, 2013
• Structured multi-class feature selection for effective face recognition, Giovanni Fusco, Luca Zini, Nicoletta Noceti, Francesca Odone, IEEE Int. Conference on Image Analysis and Processing, 2013
• what epipolar geometry can do for video-surveillance, Nicoletta Noceti, Luigi Balduzzi, Francesca Odone, IEEE Int. Conference on Image Analysis and Processing, 2013

• Action understanding
• 3D object recognition for robotics
• 3D modeling and location
• Behavior analysis
• Object detection

My research focuses on statistical learning and its application to behavior analysis problems in the field of video-surveillance tasks and automatic monitoring.

Traditional approaches based on motion analysis address many interesting applications, such as access control, anomaly detection, congestion analysis and multi-camera event description: in all these cases it is common practice to devise a measurement phase that extracts low level information from videos.

A well known limit of these methods is that while they provide effective tools to model the dynamics of a single video, they do not suffice when the problem of interest requires a higher generalization level. In the case of behaviors modeling or dynamic events classification, in particular, it is advisable to increase the abstraction of the data, designing higher-level descriptions able to model more general structures.

We may exploit the availability of possibly huge sets of examples, acquired by long time observations, and endowed with an internal structure provided by temporal coherence.

Within this framework the main goals of my research are:

• studying and developing robust methods to retrieve space-time information from videos
• studying and developing higher-level descriptions, to include space-time information within a machine learning framework
• devising machine learning strategies to model common events and anomalies from huge sets of (possibly) unlabeled examples.