The research activity has been carried out primarily on the
following fields of interest (thematic areas):
- Image and video sequence processing and analysis
1.1 Image restoration (large astronomical and medical
images)
1.2 Medical and biomedical image processing and analysis
1.3 Microscopy imaging
1.4 2-D and 3-D medical image reconstruction
1.5 Fusion of images acquired with different
devices
1.6 Video sequence processing and analysis
1.7 Capacitive image processing and analysis
1.8 3-D scene reconstruction with mobile devices
2. Pattern Recognition and machine learning
2.1 Data denoising
2.2 Feature generation and extraction
2.3 Feature analysis and reduction through using parametric
and non-parametric methods
2.4 Template matching
2.5 Classification methods
2.6 Object recognition and classification in image and sequence
2.7 Local
and global optimization methods
3. High performance computing
3.1 Parallel computing on SIMD (up to 1024 PE's) architecture
3.2 Parallel distributed computing on MIMD and hybrid SIMD/MIMD architectures
4. Computer Vision
4.1 Computer Aided Diagnosis (CAD) systems
4.1.1 Quantitative analyses of cancer perfusion by means of Computed Tomography
4.1.2 Automatic detection of regions of interest in histopathology images
4.1.3 Detection and classification of micro-calcifications
and masses in mammography
4.2
Non-invasive image-based stem cells characterization systems
4.3 Systems to measure skin aging through capacitive image analysis
4.4
Visual surveillance and monitoring systems
4.4.1 Motion detection under different environmental
conditions
4.4.2 Vehicle and people detection and tracking
4.4.3 People behaviour analysis and automatic event detection
4.5 Machine vision systems exploiting multiple 2-D and 3-D technologies
4.6 Real time systems for augmented and mixed reality
Although the research activity has been carried out on more
interdisciplinary areas, it has always kept one guiding thread. In
fact, image processing (Area 1) has been always directed towards
the automatic recognition of some image characteristics through
using pattern recognition methods (Area 2), thus often requiring
high performance (parallel or distributing) computing facilities
(Area 3) in order to achieve useful results in an acceptable amount
of time. The synergy of these three areas often converges in a
wider research field known as Computer Vision (Area 4).
