arXiv subjects

Giuseppe Patanè

Publications and source records attributed to Giuseppe Patanè.

At least 19 recordsLinked to original sources

arXiv2025

Framework of a multiscale data-driven DT of the musculoskeletal system

Musculoskeletal disorders (MSDs) are a leading cause of disability worldwide, requiring advanced diagnostic and therapeutic tools for personalised assessment and treatment. Effective management of MSDs involves the interaction of heterogeneous data sources, making the Digital Twin (DT) paradigm a valuable option. This paper introduces the Musculoskeletal Digital Twin (MS-DT), a novel framework that integrates multiscale biomechanical data with computational modelling to create a detailed, patient-specific representation of the musculoskeletal system. By combining motion capture, ultrasound imaging, electromyography, and medical imaging, the MS-DT enables the analysis of spinal kinematics, posture, and muscle function. An interactive visualisation platform provides clinicians and researchers with an intuitive interface for exploring biomechanical parameters and tracking patient-specific changes. Results demonstrate the effectiveness of MS-DT in extracting precise kinematic and dynamic tissue features, offering a comprehensive tool for monitoring spine biomechanics and rehabilitation. This framework provides high-fidelity modelling and real-time visualization to improve patient-specific diagnosis and intervention planning.

eess.IV↗

arXiv2025

3D Skin Segmentation Methods in Medical Imaging: A Comparison

Automatic segmentation of anatomical structures is critical in medical image analysis, aiding diagnostics and treatment planning. Skin segmentation plays a key role in registering and visualising multimodal imaging data. 3D skin segmentation enables applications in personalised medicine, surgical planning, and remote monitoring, offering realistic patient models for treatment simulation, procedural visualisation, and continuous condition tracking. This paper analyses and compares algorithmic and AI-driven skin segmentation approaches, emphasising key factors to consider when selecting a strategy based on data availability and application requirements. We evaluate an iterative region-growing algorithm and the TotalSegmentator, a deep learning-based approach, across different imaging modalities and anatomical regions. Our tests show that AI segmentation excels in automation but struggles with MRI due to its CT-based training, while the graphics-based method performs better for MRIs but introduces more noise. AI-driven segmentation also automates patient bed removal in CT, whereas the graphics-based method requires manual intervention.

eess.IV↗

arXiv2025

Optimal Density Functions for Weighted Convolution in Learning Models

The paper introduces the weighted convolution, a novel approach to the convolution for signals defined on regular grids (e.g., 2D images) through the application of an optimal density function to scale the contribution of neighbouring pixels based on their distance from the central pixel. This choice differs from the traditional uniform convolution, which treats all neighbouring pixels equally. Our weighted convolution can be applied to convolutional neural network problems to improve the approximation accuracy. Given a convolutional network, we define a framework to compute the optimal density function through a minimisation model. The framework separates the optimisation of the convolutional kernel weights (using stochastic gradient descent) from the optimisation of the density function (using DIRECT-L). Experimental results on a learning model for an image-to-image task (e.g., image denoising) show that the weighted convolution significantly reduces the loss (up to 53% improvement) and increases the test accuracy compared to standard convolution. While this method increases execution time by 11%, it is robust across several hyperparameters of the learning model. Future work will apply the weighted convolution to real-case 2D and 3D image convolutional learning problems.

cs.CV↗

arXiv2025

Optimal Weighted Convolution for Classification and Denosing

We introduce a novel weighted convolution operator that enhances traditional convolutional neural networks (CNNs) by integrating a spatial density function into the convolution operator. This extension enables the network to differentially weight neighbouring pixels based on their relative position to the reference pixel, improving spatial characterisation and feature extraction. The proposed operator maintains the same number of trainable parameters and is fully compatible with existing CNN architectures. Although developed for 2D image data, the framework is generalisable to signals on regular grids of arbitrary dimensions, such as 3D volumetric data or 1D time series. We propose an efficient implementation of the weighted convolution by pre-computing the density function and achieving execution times comparable to standard convolution layers. We evaluate our method on two deep learning tasks: image classification using the CIFAR-100 dataset [KH+09] and image denoising using the DIV2K dataset [AT17]. Experimental results with state-of-the-art classification (e.g., VGG [SZ15], ResNet [HZRS16]) and denoising (e.g., DnCNN [ZZC+17], NAFNet [CCZS22]) methods show that the weighted convolution improves performance with respect to standard convolution across different quantitative metrics. For example, VGG achieves an accuracy of 66.94% with weighted convolution versus 56.89% with standard convolution on the classification problem, while DnCNN improves the PSNR value from 20.17 to 22.63 on the denoising problem. All models were trained on the CINECA Leonardo cluster to reduce the execution time and improve the tuning of the density function values. The PyTorch implementation of the weighted convolution is publicly available at: https://github.com/cammarasana123/weightedConvolution2.0.

cs.CV↗

arXiv2024

US \& MRI Image Fusion Based on Markerless Skin Registration

This paper presents an innovative automatic fusion imaging system that combines 3D CT/MR images with real-time ultrasound (US) acquisition. The system eliminates the need for external physical markers and complex training, making image fusion feasible for physicians with different experience levels. The integrated system involves a portable 3D camera for patient-specific surface acquisition, an electromagnetic tracking system, and US components. The fusion algorithm comprises two main parts: skin segmentation and rigid co-registration, both integrated into the US machine. The co-registration software aligns the surface extracted from CT/MR images with patient-specific coordinates, facilitating rapid and effective fusion. Experimental testing in different settings validates the system's accuracy, computational efficiency, noise robustness, and operator independence. The co-registration error remains under the acceptable range of~$1$ cm.

cs.CV↗

arXiv2023

Learning-Based and Quality Preserving Super-Resolution of Noisy Images

Several applications require the super-resolution of noisy images and the preservation of geometrical and texture features. State-of-the-art super-resolution methods do not account for noise and generally enhance the output image's artefacts (e.g., aliasing, blurring). We propose a learning-based method that accounts for the presence of noise and preserves the properties of the input image, as measured by quantitative metrics (e.g., normalised crossed correlation, normalised mean squared error, peak-signal-to-noise-ration, structural similarity feature-based similarity, universal image quality). We train our network to up-sample a low-resolution noisy image while preserving its properties. We perform our tests on the Cineca Marconi100 cluster, at the 26th position in the top500 list. The experimental results show that our method outperforms learning-based methods, has comparable results with standard methods, preserves the properties of the input image as contours, brightness, and textures, and reduces the artefacts. As average quantitative metrics, our method has a PSNR value of 23.81 on the super-resolution of Gaussian noise images with a 2X up-sampling factor. In contrast, previous work has a PSNR value of 23.09 (standard method) and 21.78 (learning-based method). Our learning-based and quality-preserving super-resolution improves the high-resolution prediction of noisy images with respect to state-of-the-art methods with different noise types and up-sampling factors.

eess.IV↗

arXiv2023

HPC-based Solvers of Minimisation Problems for Signal Processing

Several physics and engineering applications involve the solution of a minimisation problem to compute an approximation of the input signal. Modern computing hardware and software apply high-performance computing to solve and considerably reduce the execution time. We compare and analyse different minimisation methods in terms of functional computation, convergence, execution time, and scalability properties, for the solution of two minimisation problems (i.e., approximation and denoising) with different constraints that involve computationally expensive operations. These problems are attractive due to their numerical and analytical properties, and our general analysis can be extended to most signal-processing problems. We perform our tests on the Cineca Marconi100 cluster, at the 26th position in the top500 list. Our experimental results show that PRAXIS is the best optimiser in terms of minima computation: the efficiency of the approximation is 38% with 256 processes, while the denoising has 46% with 32 processes.

cs.DC↗

arXiv2023

Graph-Based Analysis and Visualisation of Mobility Data

Urban mobility forecast and analysis can be addressed through grid-based and graph-based models. However, graph-based representations have the advantage of more realistically depicting the mobility networks and being more robust since they allow the implementation of Graph Theory machinery, enhancing the analysis and visualisation of mobility flows. We define two types of mobility graphs: Region Adjacency graphs and Origin-Destination graphs. Several node centrality metrics of graphs are applied to identify the most relevant nodes of the network in terms of graph connectivity. Additionally, the Perron vector associated with a strongly connected graph is applied to define a circulation function on the mobility graph. Such node values are visualised in the geographically embedded graphs, showing clustering patterns within the network. Since mobility graphs can be directed or undirected, we define several Graph Laplacian for both cases and show that these matrices and their spectral properties provide insightful information for network analysis. The computation of node centrality metrics and Perron-induced circulation functions for three different geographical regions demonstrate that basic elements from Graph Theory applied to mobility networks can lead to structure analysis for graphs of different connectivity, size, and orientation properties.

cs.DM↗

arXiv2023

3D Patient-specific Modelling and Characterisation of Muscle-Skeletal Districts

This work addresses the patient-specific characterisation of the morphology and pathologies of muscle-skeletal districts (e.g., wrist, spine) to support diagnostic activities and follow-up exams through the integration of morphological and tissue information. We propose different methods for the integration of morphological information, retrieved from the geometrical analysis of 3D surface models, with tissue information extracted from volume images. For the qualitative and quantitative validation, we will discuss the localisation of bone erosion sites on the wrists to monitor rheumatic diseases and the characterisation of the three functional regions of the spinal vertebrae to study the presence of osteoporotic fractures. The proposed approach supports the quantitative and visual evaluation of possible damages, surgery planning, and early diagnosis or follow-up studies. Finally, our analysis is general enough to be applied to different districts.

cs.CV↗

arXiv2023

Learning-based Framework for US Signals Super-resolution

We propose a novel deep-learning framework for super-resolution ultrasound images and videos in terms of spatial resolution and line reconstruction. We up-sample the acquired low-resolution image through a vision-based interpolation method; then, we train a learning-based model to improve the quality of the up-sampling. We qualitatively and quantitatively test our model on different anatomical districts (e.g., cardiac, obstetric) images and with different up-sampling resolutions (i.e., 2X, 4X). Our method improves the PSNR median value with respect to SOTA methods of $1.7\%$ on obstetric 2X raw images, $6.1\%$ on cardiac 2X raw images, and $4.4\%$ on abdominal raw 4X images; it also improves the number of pixels with a low prediction error of $9.0\%$ on obstetric 4X raw images, $5.2\%$ on cardiac 4X raw images, and $6.2\%$ on abdominal 4X raw images. The proposed method is then applied to the spatial super-resolution of 2D videos, by optimising the sampling of lines acquired by the probe in terms of the acquisition frequency. Our method specialises trained networks to predict the high-resolution target through the design of the network architecture and the loss function, taking into account the anatomical district and the up-sampling factor and exploiting a large ultrasound data set. The use of deep learning on large data sets overcomes the limitations of vision-based algorithms that are general and do not encode the characteristics of the data. Furthermore, the data set can be enriched with images selected by medical experts to further specialise the individual networks. Through learning and high-performance computing, our super-resolution is specialised to different anatomical districts by training multiple networks. Furthermore, the computational demand is shifted to centralised hardware resources with a real-time execution of the network's prediction on local devices.

cs.CV↗

arXiv2022

A Graphical Framework to Study the Correlation between Geometric Design and Simulation

Partial differential equations can be solved on general polygonal and polyhedral meshes, through Polytopal Element Methods (PEMs). Unfortunately, the relation between geometry and analysis is still unknown and subject to ongoing research in order to identify weaker shape-regularity criteria under which PEMs can reliably work. We propose PEMesh, a graphical framework to support the analysis of the relation between the geometric properties of polygonal meshes and the numerical performances of PEM solvers. PEMesh allows the design of polygonal meshes that increasingly stress some geometric properties, by exploiting any external PEM solver, and supports the study of the correlation between the performances of such a solver and geometric properties of the input mesh. Furthermore, it is highly modular, customisable, easy to use, and provides the possibility to export analysis results both as numerical values and graphical plots. PEMesh has a potential practical impact on ongoing and future research activities related to PEM methods, polygonal mesh generation and processing.

cs.CG↗

arXiv2022

3D Anatomical Representations and Analysis: an Application to the Spine

This work proposes a framework for the patient-specific characterization of the spine, which integrates information on the tissues with geometric information on the spine morphology. Key elements are the extraction of 3D patient-specific models of each vertebra and the intervertebral space from 3D CT images, the segmentation of each vertebra in its three functional regions, and the analysis of the tissue condition in the functional regions based on geometrical parameters. The localization of anomalies obtained in the results and the proposed visualization support the applicability of our tool for quantitative and visual evaluation of possible damages, for surgery planning, and early diagnosis or follow-up studies. Finally, we discuss the main properties of the proposed framework in terms of characterisation of the morphology and pathology of the spine on benchmarks of the spine district.

cs.GR↗

arXiv2022

A Universal Deep Learning Framework for Real-Time Denoising of Ultrasound Images

Ultrasound images are widespread in medical diagnosis for muscle-skeletal, cardiac, and obstetrical diseases, due to the efficiency and non-invasiveness of the acquisition methodology. However, ultrasound acquisition introduces noise in the signal, which corrupts the resulting image and affects further processing steps, e.g., segmentation and quantitative analysis. We define a novel deep learning framework for the real-time denoising of ultrasound images. Firstly, we compare state-of-the-art methods for denoising (e.g., spectral, low-rank methods) and select WNNM (Weighted Nuclear Norm Minimisation) as the best denoising in terms of accuracy, preservation of anatomical features, and edge enhancement. Then, we propose a tuned version of WNNM (tuned-WNNM) that improves the quality of the denoised images and extends its applicability to ultrasound images. Through a deep learning framework, the tuned-WNNM qualitatively and quantitatively replicates WNNM results in real-time. Finally, our approach is general in terms of its building blocks and parameters of the deep learning and high-performance computing framework; in fact, we can select different denoising algorithms and deep learning architectures.

eess.IV↗

arXiv2021

Secure Routine: A Routine-Based Algorithm for Drivers Identification

The introduction of Information and Communication Technology (ICT) in transportation systems leads to several advantages (efficiency of transport, mobility, traffic management). However, it may bring some drawbacks in terms of increasing security challenges, also related to human behaviour. As an example , in the last decades attempts to characterize drivers' behaviour have been mostly targeted. This paper presents Secure Routine, a paradigm that uses driver's habits to driver identification and, in particular, to distinguish the vehicle's owner from other drivers. We evaluate Secure Routine in combination with other three existing research works based on machine learning techniques. Results are measured using well-known metrics and show that Secure Routine outperforms the compared works.

cs.CR↗

arXiv2021

Fourier-based and Rational Graph Filters for Spectral Processing

Data are represented as graphs in a wide range of applications, such as Computer Vision (e.g., images) and Graphics (e.g., 3D meshes), network analysis (e.g., social networks), and bio-informatics (e.g., molecules). In this context, our overall goal is the definition of novel Fourier-based and graph filters induced by rational polynomials for graph processing, which generalise polynomial filters and the Fourier transform to non-Euclidean domains. For the efficient evaluation of discrete spectral Fourier-based and wavelet operators, we introduce a spectrum-free approach, which requires the solution of a small set of sparse, symmetric, well-conditioned linear systems and is oblivious of the evaluation of the Laplacian or kernel spectrum. Approximating arbitrary graph filters with rational polynomials provides a more accurate and numerically stable alternative with respect to polynomials. To achieve these goals, we also study the link between spectral operators, wavelets, and filtered convolution with integral operators induced by spectral kernels. According to our tests, main advantages of the proposed approach are (i) its generality with respect to the input data (e.g., graphs, 3D shapes), applications (e.g., signal reconstruction and smoothing, shape correspondence), and filters (e.g., polynomial, rational polynomial), and (ii) a spectrum-free computation with a generally low computational cost and storage overhead.

cs.LG↗

arXiv2020

Meshless Approximation and Helmholtz-Hodge Decomposition of Vector Fields

The analysis of vector fields is crucial for the understanding of several physical phenomena, such as natural events (e.g., analysis of waves), diffusive processes, electric and electromagnetic fields. While previous work has been focused mainly on the analysis of 2D or 3D vector fields on volumes or surfaces, we address the meshless analysis of a vector field defined on an arbitrary domain, without assumptions on its dimension and discretisation. The meshless approximation of the Helmholtz-Hodge decomposition of a vector field is achieved by expressing the potential of its components as a linear combination of radial basis functions and by computing the corresponding conservative, irrotational, and harmonic components as solution to a least-squares or to a differential problem. To this end, we identify the conditions on the kernel of the radial basis functions that guarantee the existence of their derivatives. Finally, we demonstrate our approach on 2D and 3D vector fields measured by sensors or generated through simulation.

cs.GR↗

arXiv2020

Continuous Fuzzy Transform as Integral Operator

The Fuzzy transform is ubiquitous in different research fields and applications, such as image and data compression, data mining, knowledge discovery, and the analysis of linguistic expressions. As a generalisation of the Fuzzy transform, we introduce the continuous Fuzzy transform and its inverse, as an integral operator induced by a kernel function. Through the relation between membership functions and integral kernels, we show that the main properties (e.g., continuity, symmetry) of the membership functions are inherited by the continuous Fuzzy transform. Then, the relation between the continuous Fuzzy transform and integral operators is used to introduce a data-driven Fuzzy transform, which encodes intrinsic information (e.g., structure, geometry, sampling density) about the input data. In this way, we avoid coarse fuzzy partitions, which group data into large clusters that do not adapt to their local behaviour, or a too dense fuzzy partition, which generally has cells that are not covered by the data, thus being redundant and resulting in a higher computational cost. To this end, the data-driven membership functions are defined by properly filtering the spectrum of the Laplace-Beltrami operator associated with the input data. Finally, we introduce the space of continuous Fuzzy transforms, which is useful for the comparison of different continuous Fuzzy transforms and for their efficient computation.

math.OC↗

arXiv2020

SceneAdapt: Scene-based domain adaptation for semantic segmentation using adversarial learning

Semantic segmentation methods have achieved outstanding performance thanks to deep learning. Nevertheless, when such algorithms are deployed to new contexts not seen during training, it is necessary to collect and label scene-specific data in order to adapt them to the new domain using fine-tuning. This process is required whenever an already installed camera is moved or a new camera is introduced in a camera network due to the different scene layouts induced by the different viewpoints. To limit the amount of additional training data to be collected, it would be ideal to train a semantic segmentation method using labeled data already available and only unlabeled data coming from the new camera. We formalize this problem as a domain adaptation task and introduce a novel dataset of urban scenes with the related semantic labels. As a first approach to address this challenging task, we propose SceneAdapt, a method for scene adaptation of semantic segmentation algorithms based on adversarial learning. Experiments and comparisons with state-of-the-art approaches to domain adaptation highlight that promising performance can be achieved using adversarial learning both when the two scenes have different but points of view, and when they comprise images of completely different scenes. To encourage research on this topic, we made our code available at our web page: https://iplab.dmi.unict.it/ParkSmartSceneAdaptation/.

cs.CV↗