Ανάπτυξη νευρωνικού δικτύου και αντίστοιχου λογισμικού εργαλείου για την αναγνώριση τρισδιάστατων αντικειμένων

Δικαίος, Χρήστος; Dikaios, Christos

dc.contributor.author	Δικαίος, Χρήστος	el
dc.contributor.author	Dikaios, Christos	en
dc.date.accessioned	2024-06-04T10:01:07Z
dc.date.available	2024-06-04T10:01:07Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/59628
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.27324
dc.rights	Default License
dc.subject	Νευρωνικά δίκτυα	el
dc.subject	Τεχνητή Νοημοσύνη	el
dc.subject	Αναγνώριση τρισδιάστατων αντικειμένων	el
dc.subject	Artificial Intelligence	en
dc.subject	3D Object Recognition	en
dc.subject	Neural Networks	en
dc.subject	You Only Look Once	en
dc.subject	Deep Learning	en
dc.title	Ανάπτυξη νευρωνικού δικτύου και αντίστοιχου λογισμικού εργαλείου για την αναγνώριση τρισδιάστατων αντικειμένων	el
dc.title	Development of an artificial neural network and corresponding software tool for 3D object recognition	en
heal.type	bachelorThesis
heal.classification	Τεχνητή Νοημοσύνη	el
heal.classification	Artificial Intelligence	en
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2023-10-02
heal.abstract	This diploma thesis aims to explore the potential of artificial intelligence and machine learning techniques, and more specifically the field of object recognition, in the maritime business. It focuses on the development of a robust and accurate system for identifying and localizing mechanical components in complex systems that can be found in a vessel, such as piping networks. The end-goal is to develop an object recognition software tool that will be user-friendly and easy to use, without the need for explicit programming and fine-tuning. The thesis reviews the state-of-the-art object recognition algorithms and analyzes their function and evolution over time. The proposed approach is based on deep learning techniques, particularly convolutional neural networks, for 3D object recognition in 2D images. Additionally, the thesis includes a method to automate the process of generating a large and accurate dataset required for training a custom object detection CNN-based network. The developed system has potential applications in various fields, including shipbuilding, manufacturing, and industrial automation, where accurate object recognition can facilitate maintenance, inspection, and retrofitting tasks.	en
heal.advisorName	Γκίνης, Αλέξανδρος	el
heal.committeeMemberName	Γκίνης, Αλέξανδρος	el
heal.committeeMemberName	Παπαδόπουλος, Χρήστος	el
heal.committeeMemberName	Βεντίκος, Νικόλαος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ναυπηγών Μηχανολόγων Μηχανικών. Τομέας Μελέτης Πλοίου και Θαλάσσιων Μεταφορών	el
heal.academicPublisherID	ntua
heal.numberOfPages	46 σ.	el
heal.fullTextAvailability	false