@InProceedings{PaulaSalv:2021:BrToRe,
author = "Paula, Davi Duarte de and Salvadeo, Denis Henrique Pinheiro",
affiliation = "{S{\~a}o Paulo State University (Unesp) - Institute of
Geosciences and Exact Sciences} and {S{\~a}o Paulo State
University (Unesp) - Institute of Geosciences and Exact
Sciences}",
title = "Breast Tomosynthesis Reconstruction Using Artificial Neural
Networks with Deep Learning",
booktitle = "Proceedings...",
year = "2021",
editor = "Paiva, Afonso and Menotti, David and Baranoski, Gladimir V. G. and
Proen{\c{c}}a, Hugo Pedro and Junior, Antonio Lopes Apolinario
and Papa, Jo{\~a}o Paulo and Pagliosa, Paulo and dos Santos,
Thiago Oliveira and e S{\'a}, Asla Medeiros and da Silveira,
Thiago Lopes Trugillo and Brazil, Emilio Vital and Ponti, Moacir
A. and Fernandes, Leandro A. F. and Avila, Sandra",
organization = "Conference on Graphics, Patterns and Images, 34. (SIBGRAPI)",
publisher = "Sociedade Brasileira de Computa{\c{c}}{\~a}o",
address = "Porto Alegre",
keywords = "Deep Learning, Breast Tomosynthesis, Image Reconstruction.",
abstract = "The Filtered Backprojection (FBP) algorithm for Computed
Tomography (CT) reconstruction can be mapped entire in an
Artificial Neural Network (ANN), with the backprojection (BP)
operation simulated analytically in a layer and the Ram-Lak filter
simulated as a convolutional layer. Thus, this work adapt the BP
layer for DBT reconstruction, making possible the use of FBP
simulated as a ANN to reconstruct DBT images. For evaluation,
Structural Similarity Index Measure (SSIM) and Peak
Signal-to-Noise Ratio (PSNR) metrics were calculated to measure
the improvement of the images made by the ANN, regarding a dataset
containing 100 virtual breast phantoms to perform the experiments.
We shown that making the Ram-Lak layer trainable, the
reconstructed image can be improved in terms of noise reduction.
And, considering an additional post-filtering step performed by
Denoising Convolutional Neural Network (DnCNN), it shown
comparable and superior results than a state-of-the-art DBT
reconstruction method, averaging 37.644 dB and 0.869 values of
PSNR and SSIM, respectively. Finally, this study enables
additional proposals of ANN with Deep Learning models for DBT
reconstruction and denoising.",
conference-location = "Gramado, RS, Brazil (virtual)",
conference-year = "18-22 Oct. 2021",
language = "en",
ibi = "8JMKD3MGPEW34M/45EHMTL",
url = "http://urlib.net/ibi/8JMKD3MGPEW34M/45EHMTL",
targetfile = "artigo_final.pdf",
urlaccessdate = "2024, May 06"
}