Publications

Recently, the field of 3D medical segmentation has been dominated by deep learning models employing Convolutional Neural Networks (CNNs) and Transformer-based architectures, each with its distinctive strengths and limitations. CNNs are constrained by a local receptive field, whereas Transformer are hindered by their substantial memory requirements as well as their data hunger, making them not ideal for processing 3D medical volumes at a fine-grained level. For these reasons, fully convolutional neural networks, as nnU-Net, still dominate the scene when segmenting medical structures in large 3D medical volumes. Despite numerous advancements toward developing transformer variants with subquadratic time and memory complexity, these models still fall short in content-based reasoning. A recent breakthrough is Mamba, a Recurrent Neural Network (RNN) based on State Space Models (SSMs), outperforming Transformers in many long-context tasks (million-length sequences) on famous natural language processing and genomic benchmarks while keeping a linear complexity. In this paper, we evaluate the effectiveness of Mamba-based architectures in comparison to state-of-the-art convolutional and Transformer-based models for 3D medical image segmentation across three well-established datasets: Synapse Abdomen, MSD BrainTumor, and ACDC. Additionally, we address the primary limitations of existing Mamba-based architectures by proposing alternative architectural designs, hence improving segmentation performances. The source code is publicly available to ensure reproducibility and facilitate further research: https://github.com/LucaLumetti/TamingMambas.

Objectives: The shape is commonly used to describe the objects. State-of-the-art algorithms in medical imaging are predominantly diverging from computer vision, where voxel grids, meshes, point clouds, and implicit surfacemodels are used. This is seen from the growing popularity of ShapeNet (51,300 models) and Princeton ModelNet (127,915 models). However, a large collection of anatomical shapes (e.g., bones, organs, vessels) and 3D models of surgical instruments is missing. Methods: We present MedShapeNet to translate datadriven vision algorithms to medical applications and to adapt state-of-the-art vision algorithms to medical problems. As a unique feature, we directly model the majority of shapes on the imaging data of real patients. We present use cases in classifying brain tumors, skull reconstructions, multi-class anatomy completion, education, and 3D printing. Results: By now, MedShapeNet includes 23 datasets with more than 100,000 shapes that are paired with annotations (ground truth). Our data is freely accessible via aweb interface and a Python application programming interface and can be used for discriminative, reconstructive, and variational benchmarks as well as various applications in virtual, augmented, or mixed reality, and 3D printing. Conclusions: MedShapeNet contains medical shapes from anatomy and surgical instruments and will continue to collect data for benchmarks and applications. The project page is: https://medshapenet.ikim.nrw/.

Negli ultimi anni, la complessità dei sistemi produttivi è aumentata significativamente a causa dei progressi nelle tecnologie derivanti dall'Industria 4.0, in particolare attraverso l'Internet of Things (IoT) e i big data. Questa evoluzione ha facilitato l'accesso senza precedenti a enormi quantità di dati, ma ha anche introdotto sfide nella raccolta dei dati e nella loro applicazione pratica per gli operatori che interagiscono con questi sistemi. Questa tesi presenta un'architettura IoT progettata per ambienti industriali reali, con l'obiettivo di dimostrare come i dati possano essere utilizzati efficacemente per monitorare le operazioni e i processi produttivi in tempo reale. L'approccio proposto migliora la capacità di rilevare e gestire i guasti, fornendo agli operatori le informazioni necessarie per prendere decisioni informate. Integrando sensori intelligenti e analisi avanzate, è possibile ottenere una visibilità dettagliata sullo stato del sistema, consentendo interventi di manutenzione tempestivi e preparando il terreno per future implementazioni di manutenzione predittiva. La ricerca include un'analisi di due casi di studio distinti, mostrando la versatilità dell'architettura in diverse applicazioni industriali. Illustra come l'utilizzo efficace dei dati possa ottimizzare l'efficienza operativa e ridurre i tempi di inattività, contribuendo così a una migliore gestione del sistema. Inoltre, questo approccio consente agli operatori umani di comprendere meglio i loro ambienti e di prendere decisioni autonome basate su informazioni in tempo reale.

Inference-time scaling has recently gained attention as an effective strategy for improving the performance of generative models without requiring additional training. Although this paradigm has been successfully applied in text and image generation tasks, its extension to video diffusion models remains relatively underexplored. Indeed, video generation presents unique challenges due to its spatiotemporal complexity, particularly in evaluating intermediate generated samples, a procedure that is required by inference-time scaling algorithms. In this work, we systematically investigate the role of the verifier: the scoring mechanism used to guide sampling. We show that current verifiers, when applied at early diffusion steps, face significant reliability challenges due to noisy samples. We further demonstrate that fine-tuning verifiers on partially denoised samples significantly improves early-stage evaluation and leads to gains in generation quality across multiple inference-time scaling algorithms, including Greedy Search, Beam Search, and a novel Successive Halving baseline.

Technologies to enable safe and effective collaboration and coexistence between humans and robots have gained significant importance in the last few years. A critical component useful for realizing this collaborative paradigm is the understanding of human and robot 3D poses using non-invasive systems. Therefore, in this paper, we propose a novel vision-based system leveraging depth data to accurately establish the 3D locations of skeleton joints. Specifically, we introduce the concept of Pose Nowcasting, denoting the capability of the proposed system to enhance its current pose estimation accuracy by jointly learning to forecast future poses. The experimental evaluation is conducted on two different datasets, providing accurate and real-time performance and confirming the validity of the proposed method on both the robotic and human scenarios.

Efficient energy provisioning is a fundamental requirement for modern transportation systems, making refueling path optimization a critical challenge. Existing solutions often focus either on inter-vehicle communication or intravehicle monitoring, leveraging Intelligent Transportation Systems, Digital Twins, and Software-Defined Internet of Vehicles with Cloud/Fog/Edge infrastructures. However, integrated frameworks that adapt dynamically to driver mobility patterns are still underdeveloped. Building on our previous PIENO framework, we present RI-PIENO (Revised and Improved Petrolfilling Itinerary Estimation aNd Optimization), a system that combines intra-vehicle sensor data with external geospatial and fuel price information, processed via IoT-enabled Cloud/Fog services. RI-PIENO models refueling as a dynamic, time-evolving directed acyclic graph that reflects both habitual daily trips and real-time vehicular inputs, transforming the system from a static recommendation tool into a continuously adaptive decision engine. We validate RI-PIENO in a daily-commute use case through realistic multi-driver, multi-week simulations, showing that it achieves significant cost savings and more efficient routing compared to previous approaches. The framework is designed to leverage emerging roadside infrastructure and V2X communication, supporting scalable deployment within next-generation IoT and vehicular networking ecosystems.

The use of self-supervised pre-training has emerged as a promising approach to enhance the performance of many different visual tasks. In this context, recent approaches have employed the Masked Image Modeling paradigm, which pre-trains a backbone by reconstructing visual tokens associated with randomly masked image patches. This masking approach, however, introduces noise into the input data during pre-training, leading to discrepancies that can impair performance during the fine-tuning phase. Furthermore, input masking neglects the dependencies between corrupted patches, increasing the inconsistencies observed in downstream fine-tuning tasks. To overcome these issues, we propose a new self-supervised pre-training approach, named Masked and Permuted Vision Transformer (MaPeT), that employs autoregressive and permuted predictions to capture intra-patch dependencies. In addition, MaPeT employs auxiliary positional information to reduce the disparity between the pre-training and fine-tuning phases. In our experiments, we employ a fair setting to ensure reliable and meaningful comparisons and conduct investigations on multiple visual tokenizers, including our proposed k-CLIP which directly employs discretized CLIP features. Our results demonstrate that MaPeT achieves competitive performance on ImageNet, compared to baselines and competitors under the same model setting. We release an implementation of our code and models at https://github.com/aimagelab/MaPeT.