Computer Vision News - May 2023

40 Congrats, Doctor Maria Chiara (cum laude)! Maria Chiara Fiorentino recently finished her PhD with the Vision Robotic and Artificial Intelligence (VRAI) group at Università Politecnica delle Marche. The aim of her thesis is to exploit Deep Learning (DL) algorithm for Ultrasound (US) image analysis. Congrats, Doctor Maria Chiara! Due to its real-time nature, Ultrasound imaging stands out among many imaging diagnostic options as a highly cost-effective modality that provides the clinician with an unrivaled and invaluable level of interaction along with other advantages (e.g. absence of ionizing radiation, portability, accessibility and cost effectiveness). Despite the well-recognized clinical utility especially for soft-tissue examination, Ultrasound presents unique challenges , such as presence of speckle and acoustic noise , which hinder image interpretation, high dependence on diagnostic operator experience and high inter- and intra-observer variability across different institutes and manufacturers’ Ultrasound systems. Deep Learning has undergone an increasing role in Ultrasound image analysis to offer decision support to clinicians. However, despite its huge potential in the Ultrasound field, there is still room for improvement. My thesis aims at moving the Deep Learning state of the art forward for already investigated Ultrasound fields, such as gynecology , a well-known and studied domain of applications; but it also aims at investigating the potentiality of Deep Learning approaches for fields, such as rheumatology , in which it has not been fully exploited. Regarding Ultrasound in gynecology, I have proposed a novel distance field network approach to automatically measure the fetal head circumference from fetal 2D Ultrasound images. Automated measurement of fetal head circumference is a crucial clinical task