Local coordinator: Gianfranco Paternò
Next_IAM is an INFN project dedicated to investigating the use of artificial intelligence methods in various medical applications.
The main field of study of the Ferrara group concerns X-ray imaging of the breast. In particular, we are involved in the development of a tool for decision support in Digital Breast Tomosynthesis (DBT) exams. From suitably pre-processed DBT slices it is possible to extract radiomic features useful for identifying any lesions using a Machine Learning algorithm. This classification can be combined with that provided by a Convolutional Neural Network (CNN), i.e. by a dedicated Deep Learning tool, to obtain a more accurate prediction of the potential presence of lesions in each slice and possibly their location.
A further field of investigation concerns the applications of AI in cardiology. In particular, the automatic analysis of plethysmographic traces, i.e., Jugular Venous Pulse (JVP) curves measured using a strain gauge applied to the neck. These curves can in fact in turn be correlated with the hemodynamic activity of the heart. An additional goal is the automatic segmentation of jugular veins and carotids from ultrasound images and then derive both JVP curves and atrial distension waveforms.