Orthopedic surgical planning depends on the ability to extract, reconstruct, and reason
over visual representations of patient anatomy. This thesis explores how modern deep
learning techniques can advance each stage of this process within an industrial context,
contributing to the development and extension of AI-driven tools for computer-aided
surgical planning. The work spans three interconnected research directions, consolidated
into four scientific contributions.
The first dir...
La presente tesi affronta i limiti degli approcci tradizionali basati su modelli per il tracciamento dei bersagli, in particolare in scenari che coinvolgono bersagli altamente manovrabili e condizioni di misurazione incerte.
Una parte significativa di questo lavoro è dedicata a un'analisi esaustiva dei metodi che integrano le reti neurali all'interno di paradigmi di stima bayesiana, evidenziandone il potenziale nel superare la rigidità dei modelli dinamici e di rumore predefiniti, pur prese...
Life, across biological systems, is characterized not by the absence of stressors but by the continuous capacity to adapt to them. Within this framework, resilience represents a dynamic, multilevel biological process that determines whether exposure to stressors results in adaptive regulation or allostatic overload. Particular attention is given to the “stress resilience signature,” conceptualized as an individual profile emerging from the dynamic interplay among endocrine, metabolic, and neu...
Climate change is reshaping agroecosystems by altering temperature and precipitation regimes, increasing the frequency of extreme events, and intensifying plant stress conditions. In maize agroecosystems these shifts not only affect plant morphophysiological traits and yield formation, but also influence plant-microbiome interactions, thereby increasing the risk of grain contamination by mycotoxins (i.e., aflatoxins and fumonisins) and potentially altering the structure and function of the rh...
Deep learning (DL) models have achieved remarkable performance across visual recogni-
tion tasks; however, their reliability under diverse and previously unseen inputs remains
a major concern. Traditional testing approaches are insufficient to systematically reveal
model failures when these models are deployed in real-world settings, as they often pro-
duce invalid inputs (i.e., outside the classification task’s domain). This thesis addresses
these challenges by advancing both the genera...
Forests and agricultural systems play a critical role in carbon sequestration, climate regulation, and sustainable resource management. However, these systems are increasingly threatened by the im-pacts of climate change, including rising temperatures, shifting precipitation patterns, and more frequent extreme weather events. These changes disrupt key ecological processes—diminishing forest carbon-storage potential and altering phenological cycles in crops such as grapevines—thereby complicat...
Il carcinoma sieroso ad alto grado dell’ovaio (HGSOC) è la neoplasia ginecologica più letale, con recidiva nel 75% delle pazienti nonostante una risposta iniziale alla chemioterapia a base di platino. L’assenza di biomarcatori predittivi affidabili rende la recidiva una sfida clinica. Il microambiente tumorale (TME), composto da cellule tumorali e non tumorali, incluse le immunitarie, gioca un ruolo cruciale nella progressione e nella risposta terapeutica. Tra le componenti immunitarie, linfo...
Glioblastoma (GBM) is the most aggressive primary malignant tumour of the central nervous system (CNS), characterized by high invasiveness, resistance to therapy and frequent recurrence. Despite the current standard treatment, including surgical resection followed by radiotherapy and chemoradiotherapy, GBM patients face poor prognosis due to therapeutic resistance and tumour recurrence. A major contributor to GBM aggressiveness is the tumour microenvironment (TME), which promotes malignancy t...
A fundamental challenge in audio software engineering is the tension between expressive, high-level programming tools and the high-performance, low-level languages required for real-time audio processing. This gap often compels developers to trade the clarity and productivity of high-level abstractions for computational efficiency. This thesis confronts these challenges through a constructive research methodology, presenting a new programming ecosystem for audio Digital Signal Processing. The...
The increasing need for sustainable refrigeration and heat pump tech nologies has driven significant research interest towards the use of carbon dioxide (CO2) as a refrigerant. Despite its environmental advantages, the transcritical behaviour of CO2 systems often leads to reduced efficiency, especially under warm climatic conditions. This doctoral research investi gates multiple strategies to enhance the energy performance, operational f lexibility, and economic viability of CO2-based systems...