Browsing by Author "Manfredi, Gilda"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Combining CNN Feature Extraction and Kolmogorov-Arnold Networks Regression for Procedural 3D Shape Generation
    (The Eurographics Association, 2025) Manfredi, Gilda; Capece, Nicola; Erra, Ugo; Gruosso, Antonio; Comino Trinidad, Marc; Mancinelli, Claudio; Maggioli, Filippo; Romanengo, Chiara; Cabiddu, Daniela; Giorgi, Daniela
    Generating 3D objects with complex, nonlinear shapes directly from images is still an open research area. To address this problem, several state-of-the-art methods use Deep Learning (DL) to predict a set of parameters from images, which are then used to generate the 3D geometry, leveraging the characteristics of procedural modeling. Recently, Kolmogorov-Arnold Networks (KANs) have emerged as an alternative to traditional Multilayer Perceptrons (MLPs) in DL, and have been successfully integrated into architectures such as Convolutional Neural Networks (CNNs), Graph Neural Networks, and Transformers. In this work, we propose a DL architecture consisting of a hybrid CNN-KAN network for parametric 3D model generation from images. The model combines the ability of KANs to capture complex nonlinear functions with the strong visual feature extraction capabilities of CNNs. The method is evaluated using both quantitative error metrics and qualitative visualizations comparing predicted shapes with ground truth, and its performance is compared against a more standard CNN-MLP architecture.
  • Thumbnail Image
    Item
    A Mixed Reality Application for Multi-Floor Building Evacuation Drills using Real-Time Pathfinding and Dynamic 3D Modeling
    (The Eurographics Association, 2024) Manfredi, Gilda; Capece, Nicola; Carlo, Rosario Pio Di; Erra, Ugo; Caputo, Ariel; Garro, Valeria; Giachetti, Andrea; Castellani, Umberto; Dulecha, Tinsae Gebrechristos
    In modern high-rise buildings, complex layouts and frequent structural changes often hinder emergency evacuation. Traditional evacuation plans, usually 2D diagrams, do not provide real-time guidance and are difficult for occupants to interpret. We propose a Mixed Reality (MR) application to address these challenges in real-time evacuation in multi-floor buildings. This application was developed on Meta Quest 3, chosen for its status as one of the best low-cost eXtended Reality (XR) headsets and a popular standalone Head-Mounted Display (HMD). Our system allows users to rapidly rescan and update building models, ensuring that evacuation guidance is always up-to-date. The proposed approach overcomes the Meta Quest 3 API's limitation of scanning only 15 rooms. It extends its capability by saving room data externally and using spatial anchors to maintain accurate alignment with the physical environment. Additionally, the application integrates Dijkstra's algorithm to dynamically calculate optimal escape routes based on the user's real-time location. A preliminary evaluation study demonstrates the application's effectiveness in enhancing situational awareness and enabling users to stay mentally sharp, highlighting its potential to improve decision-making and emergency response in dynamic building environments significantly.