Foto del docente

Samuele Salti

Professore associato

Dipartimento di Informatica - Scienza e Ingegneria

Settore scientifico disciplinare: IINF-05/A Sistemi di elaborazione delle informazioni

Didattica

Ultime tesi seguite dal docente

Tesi di Laurea

  • Realizzazione di un'interfaccia web per la valutazione della coerenza tra forme 3D e descrizioni testuali

Tesi di Laurea Magistrale

  • 3D deep learning architectures for surface classification of welded stator hairpins
  • A Comprehensive Benchmark of Efficient Text-Driven 3D Generative Models
  • A Safety Framework For Large Language Models
  • A transformer-based architecture for automated inspection of pharmaceutical labels
  • Additive decomposition of one-dimensional signals using Transformers
  • An Experimental Study on Generalization in Deep Fake Image Detection
  • Automatic Identification of Book Stamps in Old Japanese Books
  • Ball tracking in Padel Videos using Convolutional Neural Networks
  • Canonical orientation and segmentation of CT scans of skulls via 3D neural networks
  • Deep Learning on Neural Radiance Fields
  • Deepfake recognition for robust face verification
  • Diffusion Inpainting with Parameter Efficient Learning for the Generation of Architectural Elements
  • Empowering Insurance Fraud Investigations: Image Tampering Detection Strategies and Tools
  • Enhancing a Text-Shape Coherence Metric by Leveraging Contrastive Losses and a Large-Scale Dataset
  • Exploiting and Benchmarking Hardware Acceleration for Computer Vision on Industrial Embedded Devices
  • Exploring robustness to viewpoint changes by creating a dataset of simulated dashcam videos
  • Generation of Microwave Images for Sensitive Applications with Diffusion Models
  • Generative Adversarial Networks for three-dimensional multi-channel virtual fluorescent staining
  • Image Super-Resolution for Improved 6D Pose Estimation in Industrial Robotic Systems
  • Instance Segmentation of Catheters in Chest X-ray Images
  • Mitigating the effects of Severe Imbalance in Multi-class Semantic Segmentation
  • Mixing Pruning and Distillation for Lighter Diffusion Models
  • Multi-Grained Attention over Query-Scoring for Dialogue-to-Video Retrieval
  • Multi-Task Reconstruction Strategies for Unsupervised 3D Domain Adaptation
  • Physically feasible Semantic Segmentation
  • Reducing Memorization in Latent Diffusion Models for 3D Medical Images Generation
  • SkinScan - Recognition of pigmented skin lesions with Vision Transformers and Bayesian Networks
  • Test Time Training for Binary Anomaly Segmentation
  • Towards automatic photo book creation from uncurated photo collections
  • Transfer-Based Adversarial Attack Against Multimodal Models by Exploiting Perturbed Attention Region

Ultimi avvisi

Al momento non sono presenti avvisi.