Comparision Through Architectures of Semantic Segmentation in Breast Ultrasound Images Across Differents Input Data Dimensions

Clécio Silva, Salomão Mafalda, Emili Bezerra, Gustavo Oliveira de Castro, Paulo Santos Júnior, Ana Alvarez

2024

Abstract

Breast cancer is a problem that affects thousands of people every year, early diagnosis is important for the treatment of this disease. Deep learning methods shows impressive results in identification and segmentation of breast cancer task. This paper evaluates the impact of input size images on three semantic segmentation architectures applied to breast tumour ultrasound, in U-net, SegNet and DeepLabV3+. In order to (comprehensively) evaluate each architecture, 5-fold cross validation was carried out, thus reducing the impact of variations in validation and training sets. In addition, the performance of the analyzed architectures was measured using the IoU and Dice metrics. The results showed that the DeepLabV3+ architecture performed better than the others architectures in segmenting breast tumours, achieving an IoU of 0.70 and Dice of 0.73, with the input dimension of the images being 128×128.

Paper Citation

in Harvard Style

Silva C., Mafalda S., Bezerra E., Oliveira de Castro G., Santos Júnior P. and Alvarez A. (2024). Comparision Through Architectures of Semantic Segmentation in Breast Ultrasound Images Across Differents Input Data Dimensions. In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF; ISBN 978-989-758-688-0, SciTePress, pages 354-361. DOI: 10.5220/0012322500003657

in Bibtex Style

@conference{healthinf24,
author={Clécio Silva and Salomão Mafalda and Emili Bezerra and Gustavo Oliveira de Castro and Paulo Santos Júnior and Ana Alvarez},
title={Comparision Through Architectures of Semantic Segmentation in Breast Ultrasound Images Across Differents Input Data Dimensions},
booktitle={Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF},
year={2024},
pages={354-361},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012322500003657},
isbn={978-989-758-688-0},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF
TI - Comparision Through Architectures of Semantic Segmentation in Breast Ultrasound Images Across Differents Input Data Dimensions
SN - 978-989-758-688-0
AU - Silva C.
AU - Mafalda S.
AU - Bezerra E.
AU - Oliveira de Castro G.
AU - Santos Júnior P.
AU - Alvarez A.
PY - 2024
SP - 354
EP - 361
DO - 10.5220/0012322500003657
PB - SciTePress