Digital Breast Tomosynthesis (DBT) Reconstruction using Sparse Non-Local Markov Random Field (SNLMRFs) Models

• Grant number: 17/17811-2
• Support Opportunities: Scholarships abroad - Research
• Effective date (Start): March 01, 2018
• Effective date (End): February 28, 2019
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Denis Henrique Pinheiro Salvadeo
• Grantee: Denis Henrique Pinheiro Salvadeo
• Host Investigator: Andrew Douglas Arnold Maidment
• Host Institution: Instituto de Geociências e Ciências Exatas (IGCE). Universidade Estadual Paulista (UNESP). Campus de Rio Claro. Rio Claro , SP, Brazil
• Research place: University of Pennsylvania, United States

Abstract

Digital Breast Tomosynthesis (DBT), also known as 3-D Mammography, consists of a imaging system for performing mammographic exams, especially for early diagnosis of breast cancer. This imaging system was recently proposed, based on the principle of Computed Tomography (CT), but with the acquisition restricted to a small angular range (arc). Due to use dangerous ionizing radiation, researchers are constantly seeking ways to reduce the dose of ionizing radiation, but maintaining sufficient quality for medical analysis (ALARA principle). So, methods for denoising and tomographic reconstruction more suitable and robust to noise become essential for this to occur, since reducing the radiation dose of the acquisition raises the noise of the acquired image. Thus, for DBT reconstruction, this project proposes to investigate a Bayesian approach with an a priori knowledge defined by a new Markov Random Fields (MRF) model, called Sparse Non-Local MRF (SNLMRF). This new model intends to regulate the solution of the inverse problem of the reconstruction considering a triple constraint: redundancy (nonlocal information), sparsity (compressive sensing) and smoothness. It is emphasized that this model joins concepts and theories used in state-of-the-art methods of denoising. For the evaluation of the proposed approach, it should be compared with classical and state-of-the-art methods of DBT reconstruction. Finally, it is expected that the use of a more complete model (such as SNLMRF) can achieve a better balance between noise reduction and preservation of details for DBT reconstructed images, increasing the capacity for a more accurate diagnosis.

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

SCARPARO, DANIELE CRISTINA; PINHEIRO SALVADEO, DENIS HENRIQUE; GUIMARAES PEDRONETTE, DANIEL CARLOS; BARUFALDI, BRUNO; ARNOLD MAIDMENT, ANDREW DOUGLAS. Evaluation of denoising digital breast tomosynthesis data in both projection and image domains and a study of noise model on digital breast tomosynthesis image domain. JOURNAL OF MEDICAL IMAGING, v. 6, n. 3, JUL 2019. (16/09714-4, 17/17811-2, 17/25908-6)

SALVADEO, DENIS H. P.; VIMIEIRO, RODRIGO B.; VIEIRA, MARCELO A. C.; MAIDMENT, ANDREW D. A.; SCHMIDT, TG; CHEN, GH; BOSMANS, H. Bayesian Reconstruction for Digital Breast Tomosynthesis using a Non-Local Gaussian Markov Random Field a priori model. MEDICAL IMAGING 2019: PHYSICS OF MEDICAL IMAGING, v. 10948, p. 6-pg., 2019-01-01. (17/17811-2, 16/25750-0)