Precision Fishing:
Automated Detection and Species Identification of Fishes Using Deep Learning Techniques
Objectives: Development of detection algorithms for the localization and identification of marine species.
Techniques: Deep Learning (DL).
Prerequisites: Python, basic knowledge of DL preferred.
Target Audience: Master’s Level Students.
Tutor: Maria Chiara Fiorentino
Development of a vessel classification system using Time Series and AIS
Objectives: Development of a boat classification system fro small-scale fishing using Time Series (TS) and satellite images
Prerequisites: Python
Target Audience: Bachelor’s / Master’s Level Students.
Tutor: Alessandro Galdelli
Co-tutor: Adriano Mancini
Industry 4.0:
Fault Detection for ATM machines
Objectives: To develop and evaluate AI methodologies for identifying and preventing various types of failures in ATMs, with the aim of improving maintenance processes and avoiding device malfunctions. This theiss work is in collaboration with Sigma Spa, with the possibility of an internship within the company.
Techniques: Data processing, time series analytics, machine learning (ML)/deep learning
Prerequisites: Python, basic knowledge of ML preferred.
Target Audience: Master’s Level Students.
Tutor: Riccardo Rosati, Luca Romeo
Deep learning-based visual control assistant for assembly/quality control in Industry 4.0
Objectives: Develop a deep learning-based visual assistant for assembly control, enabling tool recognition and real-time assessment of tasks and actions in the production process to identify errors. Thesis work in collaboration with Sinergia Consulenze Srl.
Techniques: Deep Learning (DL), Augmented Reality/Virtual Reality (AR/VR)
Prerequisites: Python, basic knowledge of DL preferred
Target Audience: Master’s Level Students.
Tutor: Pierdicca Roberto, Riccardo Rosati
NeRF for Digital Cultural Heritage
Objectives: To use AI algorithms to create digital twins of monuments using images from drones or photos taken by users. Thesis work within the CTE Square project.
Techniques: Deep Learning (DL)
Prerequisites: Python, basic knowledge of DL preferred
Target Audience: Master’s Level Students.
Tutor: Pierdicca Roberto, Riccardo Rosati, Andrea Felicetti
Leveraging Artificial Intelligence to understand street life in real-time
Objectives: To use AI algorithms to analyze and understand urban life in real-time, focusing on street dunamics. This can have applications in various sectors, such as public safety, urban planning or traffic management. Thesis work within the CTE Square project.
Techniques: Deep Learning (DL)
Prerequisites: Python, basic knowledge of DL preferred
Target Audience: Master’s Level Students.
Tutor: Riccardo Rosati, Andrea Felicetti Daniele Berardini
Retail:
Retail Shelf segmentation/detection
Objectives: Use of lightweight detection/segmentation models to detect different shelves from a photo.
Techniques: Deep Learning (DL)
Prerequisites: Python, basic knowledge of DL preferred
Target Audience: Master’s Level Students.
Tutor: Rocco Pietrini
LLMs for Retail product identification and description
Objectives: Use of Large Language Models (LLMs) for the description and extraction of information from retail product images.
Techniques: Deep Learning (DL)
Prerequisites: Python, basic knowledge of DL preferred
Target Audience: Bachelor’s Level Students.
Tutor: Rocco Pietrini
Synthetic – Real image matching in retail
Objectives: Matching of real images and synthetic models for retail products (GAN, embedding vectors).
Techniques: Deep Learning (DL)
Prerequisites: Python, basic knowledge of DL preferred
Target Audience: Master’s Level Students.
Tutor: Rocco Pietrini
Embedding networks explainability
Objectives: Explainability on embedding vectors for image matching
Techniques: Deep Learning (DL)
Prerequisites: Python, basic knowledge of DL preferred
Target Audience: Master’s Level Students.
Tutor: Rocco Pietrini