Cookies Policy
We use cookies to improve our site and your experience. By continuing to browse our site you accept our cookie policy. Find out More
Close
  • Menu
About

About

I received the MSc. in Bioengineering - Biomedical Engineering,  from the Faculty of Engineering of University of Porto (FEUP), Portugal, in March, 2016.

Early in my undergraduate years, I interned at National Institute of Biomedical Engineering (INEB), contributing in the NeurOn project - a project founded by FCT to conceive a new approach for regeneration and functional recovery of spinal cord injury. Here, I discovered my interest of promoting research through the use image analysis as a tool to obtain results in a faster, more reliable and less subjective way.

The leapt for medical image analysis came with the opportunity to join the Breast Cancer Survey (BCS) project, a project in a consortium of FEUP, Coimbra University (UC), Portuguese League Against Cancer (LPCC) and Emílio Azevedo Campos.SA (EAC.SA),  which aimed at developing automatic modules of screening and diagnosis of breast cancer to be implemented on PACS.

Currently, I am a PhD student in the informatics joint programme MAPi, a member of Visual Computing and Machine Learning (VCMI) and Breast Research Groups at INESC TEC, and a collaborator in the project BCCT.plan.

My main research interests are image analysis, machine learning and decision support systems, particularly in the area of breast cancer. 

Interest
Topics
Details

Details

Publications

2018

A Regression Model for Predicting Shape Deformation after Breast Conserving Surgery

Authors
Zolfagharnasab, H; Bessa, S; Oliveira, SP; Faria, P; Teixeira, JF; Cardoso, JS; Oliveira, HP;

Publication
Sensors

Abstract

2017

Prediction of Breast Deformities: A Step Forward for Planning Aesthetic Results After Breast Surgery

Authors
Bessa, S; Zolfagharnasab, H; Pereira, E; Oliveira, HP;

Publication
Pattern Recognition and Image Analysis - 8th Iberian Conference, IbPRIA 2017, Faro, Portugal, June 20-23, 2017, Proceedings

Abstract
The development of a three-dimensional (3D) planing tool for breast cancer surgery requires the existence of proper deformable models of the breast, with parameters that can be manipulated to obtain the desired shape. However, modelling breast is a challenging task due to the lack of physical landmarks that remain unchanged after deformation. In this paper, the fitting of a 3D point cloud of the breast to a parametric model suitable for surgery planning is investigated. Regression techniques were used to learn breast deformation functions from exemplar data, resulting in comprehensive models easy to manipulate by surgeons. New breast shapes are modelled by varying the type and degree of deformation of three common deformations: ptosis, turn and top-shape. © Springer International Publishing AG 2017.

2017

Registration of Breast Surface Data Before and After Surgical Intervention

Authors
Bessa, S; Oliveira, HP;

Publication
Pattern Recognition and Image Analysis - 8th Iberian Conference, IbPRIA 2017, Faro, Portugal, June 20-23, 2017, Proceedings

Abstract
Surgery planing of breast cancer interventions is gaining importance among physicians, who recognize value in discussing the possible aesthetic outcomes of surgery with patients. Research is been propelled to create patient-specific breast models, but breast image registration algorithms are still limited, particularly for the purpose of matching pre- and post-surgical data of patient’s breast surfaces. Yet, this is a fundamental task to learn prediction models of breast healing process after surgery. In this paper, a coarse-to-fine registration strategy is proposed to match breast surface data acquired before and after surgery. Methods are evaluated in their ability to register surfaces in an anatomical reliable way, and results suggest proper alignment adequated to be used as input to train deformable models. © Springer International Publishing AG 2017.

2014

Normal breast identification in screening mammography: A study on 18 000 images

Authors
Bessa, S; Domingues, I; Cardoso, JS; Passarinho, P; Cardoso, P; Rodrigues, V; Lage, F;

Publication
Proceedings - 2014 IEEE International Conference on Bioinformatics and Biomedicine, IEEE BIBM 2014

Abstract
Through the years, several CAD systems have been developed to help radiologists in the hard task of detecting signs of cancer in the numerous screening mammograms. A more recent trend includes the development of pre-CAD systems aiming at identifying normal mammograms instead of detecting suspicious ones. Normal breasts are screened-out from the process, leaving radiologists more time to focus on more difficult cases. In this work, a new approach for the identification of normal breasts is presented. Considering that even breasts with malignant findings are mostly constituted by normal tissue, the breast area is divided into blocks which are then compared pairwise. If all blocks are very similar, the breast is labelled as normal, and as suspicious otherwise. Features characterizing the pairwise block similarity and characterizing the intra-block pixel distribution are used to design a predictive method based on machine learning techniques. The proposed solution was applied on a real world screening setting composed by nearly 18000 mammograms. Results are similar to the more complex state of the art approaches by correctly identifying more than 20% of the normal mammograms. These results suggest the usefulness of the relative comparison instead of the absolute classification. When properly used, simple statistics can suffice to distinguish the clearly normal breasts. © 2014 IEEE.

2013

Automatic quantification of cell outgrowth from neurospheres

Authors
Bessa, S; Quelhas, P; Amaral, IF;

Publication
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Abstract
In the development of new regenerative medicine therapies for the treatment of central nervous system and spinal cord injuries, the identification of factors that inhibit or promote cell outgrowth in neurite outgrowth assays is fundamental, and the neurotrophic activity is commonly assessed based on the neurite/cell outgrowth. Neurites are projections from the cell body or the initial neurosphere and typically present low-contrast to background in phase contrast images. The extent of neurites is usually measured in a manual way and fluorescence images are the most used, generally requiring imunofluorescent staining. We present a novel automatic approach for the quantification of cell outgrowth from neurospheres, based on phase contrast and fluorescence images acquired from samples merely processed for DNA staining. Our approach detects the neurite/cell outgrowth, and its measures are in high agreement with the ones obtained manually. Furthermore, the image analysis time was reduced in more than 95% allowing the increase of the amount of data to be analyzed. © 2013 Springer-Verlag.

Supervised
thesis

2017

Multimodal Medical Image Registration: creating a complete 3D model of the women breast

Author
Sílvia da Conceição Neto Bessa

Institution
UP-FEUP

2017

Digital image colorimetry for determination of sulfonamides in water

Author
Paulo Jorge Teixeira Silva

Institution
UP-FEUP

2017

Breast Modelling Towards an Educational Tool for Breast Cancer Surgeons

Author
Daniela do Vale Afonso

Institution
UP-FEUP