We are pleased to welcome collaborations with researchers and students interested in the mission of our laboratory.
Please Contact us if you are interested in the following or related topics:
Medical Image Processing and Analysis
Medical image processing and analysis is a complex branch of science that lies at the intersection of applied mathematics,
computer science, physics, statistics, and biomedical sciences. In the past years, many techniques had
been developed such as segmentation and texture analysis, which were used for cancer and other disorder identifications.
In addition, image registering and fusion methods had been developed to address new modalities such as PET-CT.
The use of external knowledge about the segmented regions of interest allows for powerful image analysis algorithm development.
Radiomics: identification of quantitative biomarkers for image characterization
Radiomics is defined as the high-throughput extraction and analysis of large amounts
of quantitative features from medical imaging data to characterize tumor phenotypes.
The pre-processed and selected features are used for the development of efficient AI models
for image analysis and classification. In current years the laboratory is developing the Radiomics of breast
lesions on MRI and Mammographic images using also Machine Learning and Deep Learning algorithms.
Geometric Algebra for Medical Imaging
Geometric Algebra (GA) offers a powerful mathematical and computational model
for the development of effective medical imaging algorithms. The practical use of GA-based methods in medical
imaging requires fast and efficient implementations to meet real-time processing constraints as well as accuracy
and robustness requirements. The purpose of the research line is to develop state of the art GA-based techniques
for medical image analysis and processing.
Development of Explainable and Interpretable AI systems
In recent years, AI models trained through Machine Learning and Deep Learning algorithms,
have achieved very promising performances in many contexts dealing with an high numbers of annotated examples.
However, their use in a real clinical practice and environment is quite limited due to their 'black-box' nature.
Trusting a medical system returning yes/no, malignant/benign, lesion/not lesion is very hard: an explanation of
the output is required. In our laboratory, model development is supported by Explainable AI algorithms in order
to make the models transparent and give an explanation of the results obtained.
Development of Clinical Decision Support Systems
Artificial intelligence systems for the diagnosis or prediction of diseases aim to support
doctor's work and never replace it. For this reason, in our laboratory, the close interaction between doctors
and engineers enables efficient tools and plug-ins development to be integrated in already consolidated
image analysis platforms in real clinical practice.
Rapid embedded sensors prototyping
Medical imaging involves important computational problems in image segmentation and analysis,
shape approximation, three-dimensional (3D) modeling, and registration of volumetric data. Specialized co-processing
architectures offering direct hardware support are needed to achieve real-time performance. In our laboratory, several
embedded architectures have been prototyped on complete System-on-Programmable-Chip (SoPC) boards and devices.
Experimental results show average speedups of one order of magnitude for the above reported algorithms.
Dry Lab and surgery rooms
Dry Labs are designed primarily to perform any kind of computational or applied mathematics
to solve complex problems linked to the surgery practice. Several studies have demonstrated its effectiveness as a
training tool to improve physician performance in operation rooms. The research line aims to develop and use Virtual
and Augmented Reality solutions, 3D model reconstruction techniques and 3D printing devices to develop a cross link
between Dry Labs and Operating Rooms.