@InProceedings{Pechlivanis_US-DICOMizer_2026,
author="Pechlivanis, Dimitrios and Didaskalou, Stylianos and Kaldoudi, Eleni and Drosatos, George",
editor="Fred, Ana and Gamboa, Hugo and Yousef, Malik and Bhavsar, Arnav and Pelayo, Sylvia and Kim, Jungsil and Concei{\c{c}}{\~a}o, Raquel",
title="US-DICOMizer: An Open-Source Tool for Preparing Ultrasound DICOM Data for AI Applications",
booktitle="Biomedical Engineering Systems and Technologies",
year="2026",
publisher="Springer Nature Switzerland",
address="Cham",
pages="502--521",
series="Communications in Computer and Information Science (CCIS)",
volume="3115",
doi="10.1007/978-3-032-34459-5_24",
abstract="Ultrasound imaging is a widely used diagnostic modality in diverse clinical contexts, but preparing ultrasound data for artificial intelligence (AI) applications presents notable challenges -- including the need to ensure data privacy, extract diagnostically relevant regions, and associate contextual metadata. This paper presents US-DICOMizer, a standalone, open-source application designed to streamline the preparation of ultrasound DICOM files for AI workflows. The tool supports three core functions: (1) anonymisation, to remove sensitive metadata in compliance with privacy regulations; (2) cropping, to isolate relevant regions in images or multiframe videos; and (3) tagging, to enrich each file with metadata such as anatomical position and imaging purpose. Built with a user-friendly interface and a robust processing backend, US-DICOMizer incorporates semi-automated features, including a newly introduced autocropping function, to enhance workflow efficiency. Its effectiveness was evaluated using deep vein thrombosis (DVT) ultrasound datasets, showing fast processing times and favourable usability feedback from 15 healthcare professionals. These results demonstrate the tool's suitability for scalable, standardised image preparation in AI development pipelines. The application offers a generalisable framework for ultrasound data preparation, while being adaptable to broader clinical use cases. Future work will focus on expanding support to other imaging modalities, integrating AI-assisted features, and conducting large-scale clinical validations.",
isbn="978-3-032-34459-5"
}