Cookies Policy
The website need some cookies and similar means to function. If you permit us, we will use those means to collect data on your visits for aggregated statistics to improve our service. Find out More
Accept Reject
  • Menu
About
Download Photo HD

About

Tomé Albuquerque received a B.Sc. degree in biochemistry from University of Aveiro and a M.S. degree in biomedical engineering from University of Porto. He is currently pursuing the Ph.D. degree with the University of Porto. He is also a Researcher at INESC TEC since 2019. His main research interests include machine learning, computer vision, and medical imaging diagnosis.

Interest
Topics
Details

Details

001
Publications

2021

Ordinal losses for classification of cervical cancer risk

Authors
Albuquerque, T; Cruz, R; Cardoso, JS;

Publication
PEERJ COMPUTER SCIENCE

Abstract
Cervical cancer is the fourth leading cause of cancer-related deaths in women, especially in low to middle-income countries. Despite the outburst of recent scientific advances, there is no totally effective treatment, especially when diagnosed in an advanced stage. Screening tests, such as cytology or colposcopy, have been responsible for a substantial decrease in cervical cancer deaths. Cervical cancer automatic screening via Pap smear is a highly valuable cell imaging-based detection tool, where cells must be classified as being within one of a multitude of ordinal classes, ranging from abnormal to normal. Current approaches to ordinal inference for neural networks are found to not sufficiently take advantage of the ordinal problem or to be too uncompromising. A non-parametric ordinal loss for neuronal networks is proposed that promotes the output probabilities to follow a unimodal distribution. This is done by imposing a set of different constraints over all pairs of consecutive labels which allows for a more flexible decision boundary relative to approaches from the literature. Our proposed loss is contrasted against other methods from the literature by using a plethora of deep architectures. A first conclusion is the benefit of using non-parametric ordinal losses against parametric losses in cervical cancer risk prediction. Additionally, the proposed loss is found to be the top-performer in several cases. The best performing model scores an accuracy of 75.6% for seven classes and 81.3% for four classes.

2021

Embedded regularization for classification of colposcopic images

Authors
Albuquerque, T; Cardoso, JS;

Publication
Proceedings - International Symposium on Biomedical Imaging

Abstract
Cervical cancer ranks as the fourth most common cancer among females worldwide with roughly 528, 000 new cases yearly. Significant progress in the realm of artificial intelligence particularly in neural networks and deep learning help physicians to diagnose cervical cancer more accurately. In this paper, we address a classification problem with the widely used VGG16 architecture. In addition to classification error, our model considers a regularization part during tuning of the weights, acting as prior knowledge of the colposcopic image. This embedded regularization approach, using a 2D Gaussian kernel, has enabled the model to learn which sections of the medical images are more crucial for the classification task. The experimental results show an improvement compared with standard transfer learning and multimodal approaches of cervical cancer classification in literature. © 2021 IEEE.

2019

Analysing the peripheral nerve tissue using distinct discretization techniques

Authors
Albuquerque, TM; Belinha, J; Natal Jorge, RMN;

Publication
2019 6TH IEEE PORTUGUESE MEETING IN BIOENGINEERING (ENBENG)

Abstract
The main purpose of this work is to study the biomechanical behaviour of a peripheral nerve tissue when submitted to mechanical stretching forces. Nerve injury reduces life quality and has a strong impact on the national productivity, so it is essential to study the biomechanical properties of peripheral nerves to improve the repair and regeneration procedure. The study was conducted on a 2D and a 3D model of a sciatic nerve bifurcation in the lower thigh. These models were inspired on a real nerve bifurcation and have been developed using several computational tools. It was used the finite element method (FEM), Radial Point Interpolation Method (RPIM) and Neighbour Radial Point Interpolation Method (NNRPIM) to perform the analyses in FEMAS (R). The results of the analysis with the three different methods are very similar, the main stress is observed always in the same region in both 2D and 3D models and the displacement results for the selected points in the two models are concordant. The results obtained by the three different analysis methods are very similar which not only allow to conclude that these methods are appropriate numerical tools to analyse the biomechanical behaviour of peripheral nerve tissue but also confirms the robustness of the used methods.