View in EDS

Sampling type operators versus AI in medical image processing

Saved in:
Bibliographic Details
Title: Sampling type operators versus AI in medical image processing
Authors: Rinelli, Lucrezia, Travaglini, Arianna, Vescera, Nicolò, Vinti, Gianluca
Contributors: Rinelli, Lucrezia, Travaglini, Arianna, Vescera, Nicolò, Vinti, Gianluca
Publication Year: 2025
Collection: IRIS Università degli Studi di Perugia
Subject Terms: Sampling Kantorovich operators, convergence results, segmentation algorithm, image processing, U-Net neural network, similarity indices
Description: This paper presents a study that leverages the important and innovative approximation properties of sampling Kantorovich operators, implemented in the processing of computed tomography (CT) images of patients with abdominal aortic aneurysm (AAA). By exploiting the remarkable reconstruction capabilities of these operators, we investigate a deterministic model grounded in a segmentation algorithm where sampling Kantorovich operators play a crucial role. This mathematically-based approach is compared with an artificial intelligence-based method employing a U-Net neural network. Both methods aim to segment the patent area of the aortic vessel, offering innovative and alternative techniques to the nephrotoxic contrast agents, typically used in diagnosing AAA. The results obtained from testing both methods were evaluated numerically and visually, demonstrating that both approaches yield accurate outcomes.
Document Type: article in journal/newspaper
Language: English
Relation: info:eu-repo/semantics/altIdentifier/wos/WOS:001593290900001; volume:0; issue:0; firstpage:0; lastpage:0; numberofpages:1; journal:DISCRETE AND CONTINUOUS DYNAMICAL SYSTEMS. SERIES S; https://hdl.handle.net/11391/1608335
DOI: 10.3934/dcdss.2025147
Availability: https://hdl.handle.net/11391/1608335
https://doi.org/10.3934/dcdss.2025147
https://www.aimsciences.org/article/doi/10.3934/dcdss.2025147
Accession Number: edsbas.99569076
Database: BASE
Description
Abstract:This paper presents a study that leverages the important and innovative approximation properties of sampling Kantorovich operators, implemented in the processing of computed tomography (CT) images of patients with abdominal aortic aneurysm (AAA). By exploiting the remarkable reconstruction capabilities of these operators, we investigate a deterministic model grounded in a segmentation algorithm where sampling Kantorovich operators play a crucial role. This mathematically-based approach is compared with an artificial intelligence-based method employing a U-Net neural network. Both methods aim to segment the patent area of the aortic vessel, offering innovative and alternative techniques to the nephrotoxic contrast agents, typically used in diagnosing AAA. The results obtained from testing both methods were evaluated numerically and visually, demonstrating that both approaches yield accurate outcomes.
DOI:10.3934/dcdss.2025147