Analysis of vertebrae without fracture on spine MRI to assess bone fragility: A Comparison of Traditional Machine Learning and Deep Learning
Ramos, Jonathan S. and de Aguiar, Erikson J. and Belizario, Ivar V. and Costa, Márcus V. L. and Maciel, Jamilly G. and Cazzolato, Mirela T. and Traina, Caetano and Nogueira-Barbosa, Marcello H. and Traina, Agma J. M.
2022 IEEE 35th International Symposium on Computer-Based Medical Systems (CBMS).
[bibtex] [doi: 10.1049/ipr2.12242]

@INPROCEEDINGS{9867008,
author={Ramos, Jonathan S. and de Aguiar, Erikson J. and
Belizario, Ivar V. and Costa, Márcus V. L. and Maciel, Jamilly G. and
Cazzolato, Mirela T. and Traina, Caetano and Nogueira-Barbosa, Marcello H. and
Traina, Agma J. M.},
booktitle={2022 IEEE 35th International Symposium on Computer-Based
Medical Systems (CBMS)}, 
title={Analysis of vertebrae without fracture on spine MRI to assess bone
fragility: A Comparison of Traditional Machine Learning and Deep Learning}, 
year={2022},
volume={},
number={},
pages={78-83},
keywords={Deep learning;Support vector machines;Osteoporosis;Sensitivity;
Databases;Magnetic resonance imaging;Magnetic resonance;
Magnetic resonance imaging;machine learning;deep learning;
vertebral fragility fractures;texture analysis},
doi={10.1109/CBMS55023.2022.00021},
abstract = {Bone mineral density (BMD) is the international standard
for evaluating osteoporosis/osteopenia. The success rate of BMD alone
in estimating the risk of vertebral fragility fracture (VFF) is
approximately 50%, making BMD far from ideal in predicting VFF.
In addition, whether or not a patient has been diagnosed with
osteoporosis or osteopenia, he or she may suffer a VFF. For
this reason, we conducted an extensive empirical study to
assess VFFs in postmenopausal women. We considered a representative
dataset of 94 T1- and T2-weighted routine spine MRI (with osteopenia
or osteoporosis), split into 2,400 samples (slices). Comparing the
classification results of machine learning and deep learning (DL)
techniques showed that DL generally achieved better results at
the cost of higher computational power and hard explainability.
ResNet achieved the best results in discriminating patients from
groups with and without VFFs with 83% accuracy and 90% AUC (with
a confidence interval of 99%). Our results represent a significant
step toward prospective and longitudinal studies investigating
methods to achieve higher accuracy in predicting VFFs based
on spine MRI features of vertebrae without fracture.}
}

Automatic image segmentation based on label propagation
Ivar Vargas Belizario, Oscar Cuadros Linares, João do Espirito Santo Batista Neto
IET Image Processing, Volume 15, Issue 11, September 2021.
[bibtex] [doi: 10.1049/ipr2.12242]

@article{https://doi.org/10.1049/ipr2.12242,
author = {Belizario, Ivar Vargas and Linares, Oscar Cuadros and Neto,
João do Espirito Santo Batista},
title = {Automatic image segmentation based on label propagation},
journal = {IET Image Processing},
volume = {15},
number = {11},
pages = {2532-2547},
doi = {https://doi.org/10.1049/ipr2.12242},
url = {https://ietresearch.onlinelibrary.wiley.com/doi/abs/10.1049/ipr2.12242},
eprint = {https://ietresearch.onlinelibrary.wiley.com/doi/pdf/10.1049/ipr2.12242},
abstract = {Abstract This article introduces an automatic approach for the
segmentation of coloured natural scene images based on graphs and the
propagation of labels originally designed for communities detection in
complex networks. Images are initially pre-segmented with super-pixels,
followed by feature extraction using colour information of each super-pixels.
The resulting graph consists of vertices which represent super-pixels,
whereas the edge weights are a measure of similarity between super-pixels.
The resulting segmentation corresponds to the propagation of labels among
the vertices. In this article, three strategies for propagating labels have
been formulated: (i) iterative propagation (ILP), (ii) recursive 
propagation (RLP) and (iii) a weighted recursive propagation (WRLP).
The experiments have shown that the proposed methods, when compared to
other state-of-the-art methods, produce better results in terms of
segmentation quality and processing time.},
year = {2021}
}

Graphs from Features: Tree-Based Graph Layout for Feature Analysis
Rosane Minghim 1, Liz Huancapaza, Erasmo Artur, Guilherme P. Telles, Ivar V. Belizario
Algorithms, Graph Drawing and Information Visualization, Volume 13, Issue 11, November 2020
[bibtex] [doi: 10.3390/a13110302]

@Article{rosane2020,
AUTHOR = {Minghim, Rosane and Huancapaza, Liz and Artur, Erasmo and 
Telles, Guilherme P. and Belizario, Ivar V.},
TITLE = {Graphs from Features: Tree-Based Graph Layout for Feature Analysis},
JOURNAL = {Algorithms},
VOLUME = {13},
YEAR = {2020},
NUMBER = {11},
ARTICLE-NUMBER = {302},
URL = {https://www.mdpi.com/1999-4893/13/11/302},
ISSN = {1999-4893},
ABSTRACT = {Feature Analysis has become a very critical task in data
analysis and visualization. Graph structures are very flexible in
terms of representation and may encode important information on
features but are challenging in regards to layout being adequate
for analysis tasks. In this study, we propose and develop
similarity-based graph layouts with the purpose of locating relevant
patterns in sets of features, thus supporting feature analysis and
selection. We apply a tree layout in the first step of the strategy,
to accomplish node placement and overview based on feature similarity.
By drawing the remainder of the graph edges on demand, further
grouping and relationships among features are revealed. We evaluate
those groups and relationships in terms of their effectiveness
in exploring feature sets for data analysis. Correlation of features
with a target categorical attribute and feature ranking are added
to support the task. Multidimensional projections are employed to
plot the dataset based on selected attributes to reveal the 
effectiveness of the feature set. Our results have shown that the
tree-graph layout framework allows for a number of observations that
are very important in user-centric feature selection, and not easy
to observe by any other available tool. They provide a way of finding
relevant and irrelevant features, spurious sets of noisy features,
groups of similar features, and opposite features, all of which are
essential tasks in different scenarios of data analysis. Case studies
in application areas centered on documents, images and sound data
demonstrate the ability of the framework to quickly reach a
satisfactory compact representation from a larger feature set.},
DOI = {10.3390/a13110302}
}

Multi-level Graph Label Propagation for Image Segmentation
Ivar Vargas Belizario, Joao Batista Neto
33rd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI), November 2020
[bibtex] [doi: 10.1109/SIBGRAPI51738.2020.00034]

@INPROCEEDINGS{9265998,
author={Belizario, Ivar Vargas and Neto, Joao Batista},
booktitle={2020 33rd SIBGRAPI Conference on Graphics, Patterns
and Images (SIBGRAPI)}, 
title={Multi-level Graph Label Propagation for Image Segmentation}, 
year={2020},
volume={},
number={},
pages={195-202},
abstract = {
This article introduces a multi-level automatic image segmentation
method based on graphs and Label Propagation (LP), originally
proposed for the detection of communities in complex networks,
namely MGLP. To reduce the number of graph nodes, a super-pixel
strategy is employed, followed by the computation of color
descriptors. Segmentation is achieved by a deterministic
propagation of vertex labels at each level. Several experiments
with real color images of the BSDS500 dataset were performed to
evaluate the method. Our method outperforms related strategies
in terms of segmentation quality and processing time. Considering
the Covering metric for image segmentation quality, for example,
MGLP outperforms LPCI-SP, its most similar counterpart, in 38.99%.
In term of processing times, MGLP is 1.07 faster than LPCI-SP.}
doi={10.1109/SIBGRAPI51738.2020.00034}}