FFFabio Augusto Faria

Fabio Augusto Faria

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Fabio Augusto Faria has a total of 63 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Citizen-science-supported Sentinel-2 deforestation monitoring: asymmetric geographic transfer across contrasting Amazonian frontiers

    by Juan Carlos Guerra, Hugo Resende, Fabio Faria, Alvaro Luiz Fazenda

    Operational deforestation monitoring underpins conservation and forest governance, but its scalability depends on reliable and geographically representative training data. We evaluated the geographic transferability of the ForestEyes citizen-science framework across two contrasting deforestation frontiers in the Peruvian Amazon using Sentinel-2 imagery from nine lowland Regions of Interest in mining-dominated Madre de Dios and agriculture-dominated Ucayali. Linear Support Vector Machine classifiers were trained using either GeoBosques reference labels or ForestEyes majority-vote labels and evaluated under bidirectional within- and cross-domain transfer.We also tested whether Shannon-entropy-based sample selection improved transfer performance and whether volunteer-response uncertainty was associated with segment and landscape characteristics. ForestEyes labels showed 92.2% overall agreement with GeoBosques. Geographic transfer was strongly asymmetric: mining-to-agriculture balanced accuracy reached 88.6% with GeoBosques labels and 85.2% with ForestEyes labels, whereas agriculture-to-mining transfer reached 61.3% and 76.1%, respectively. Median-entropy filtering increased reference agreement from 92.2% to 96.0% in the retained subset but did not consistently improve model performance.Higher segment homogeneity was associated with lower volunteer-response uncertainty, whereas landscape-context variables showed weaker and less consistent associations. These findings show that citizen-science labels can support operational deforestation monitoring across contrasting Amazonian frontiers, while revealing that geographic representativeness, not only label confidence or segment quality, is critical for reliable model transfer between mining- and agriculture-dominated landscapes.

  • Open Access

    Random‐key optimization for 2D irregular packing with reusable area evaluation

    by Felipe S. C. Roberto, Victor U. Pugliese, Sanderson L. Gonzaga de Oliveira, Oseias Ferreira, Antonio A. Chaves, Fabio A. Faria

    The diverse constraints of industrial applications lead to variants of two‐dimensional (2D) irregular packing problems that require tailored solution methods. This paper addresses a real‐world industrial challenge by proposing a new problem definition, the maximum reusable contiguous area problem (MRCAP), and a novel metric, the maximum contiguous area, to measure and maximize the contiguous unused area in a layout, thereby facilitating the reuse of remnant material. This study proposes an approach focused on optimizing placement policies. We develop a decoder, implemented within a new version of the random‐key optimizer (RKO) framework, that dynamically assigns the best placement rule from an 11‐heuristic portfolio. We validate our methodology on established literature benchmarks. Among the 15 benchmark 2D Irregular Knapsack Problem instances evaluated, RKO matched the best‐performing existing algorithm on 11 and achieved better solutions on the remaining 4. These results suggest that RKO is competitive with, and potentially superior to, this algorithm. RKO's results indicate that current benchmarks no longer adequately represent the problem, motivating the introduction of extended benchmark instances. RKO also outperforms leading 2D irregular strip packing problem (SPP) methods relying on constructive sequence search. Finally, comparing our method against the top SPP algorithm, which minimizes layout overlap, on real‐world MRCAP instances shows that RKO yields superior remnant quality in almost all problem cases. This demonstrates that minimizing layout width in SPP does not ensure effective remnant valorization, highlighting the distinction between these two optimization objectives.


  • FaceMixup: Enhancing Facial Expression Recognition through Mixed Face Regularization

    by Mateus Souza, Fabio A. Faria, Raoni Texeira, Mauricio P. Segundo

  • Reliability-Aware Citizen Science for Environmental Machine Learning

    by Hugo Resende, Eduardo B. Neto, Fabio A. Cappabianco, Álvaro L. Fazenda, Fabio A. Faria

  • Do Superpixel Segmentation Methods Influence Deforestation Image Classification?

    by Hugo Resende, Fabio A. Faria, Eduardo B. Neto, Isabela Borlido, Victor Sundermann, Silvio Jamil F. Guimarães, Álvaro L. Fazenda

  • Open Access

    Crowd-Powered Sampling for Machine Learning: Leveraging Citizen Scientist Response Patterns in AutoML Workflows

    by Hugo Resende, Eduardo B. Neto, Fabio A. M. Cappabianco, Álvaro L. Fazenda, Fabio A. Faria

    Defining effective models for data classification is challenging, especially in complex contexts. Automated Machine Learning (AutoML) tools can assist in this process by generating rankings tailored to the nature of the data and the problem. In this work, we investigate the performance of five classifiers applied to the task of deforestation segment classification, using data labeled through a citizen science campaign from the ForestEyes project. We selected SVM, Ridge, AdaBoost, KNN, and MLP models based on a ranking generated with the PyCaret AutoML library, prioritizing diverse modeling approaches. Initially, the performance of the models is assessed using the incremental training strategy based on entropy of the volunteer's classifications. Then, a new training strategy is proposed based on the median response time of volunteers when evaluating each segment, exploring three ordering strategies: ascending, descending, and edge-based. Experimental results aligned with the PyCaret ranking, with SVM achieving the best performance, followed by Ridge and AdaBoost, especially when trained on smaller and more reliable data subsets. Both the entropy-based approach and the new strategy using median response time demonstrated strong potential to efficiently train machine learning models in scenarios with scarce data, typical in citizen science campaigns.


  • Open Access

    Citizen Science in Action: Building Trust in Deforestation Data Through Continuous Reliability Enhancements

    by Hugo Resende, Álvaro L. Fazenda, Amanda A. S. de Oliveira, Fábio A. M. Cappabianco, Fabio A. Faria

    The reliability of data produced by citizen science projects is crucial for applications that depend on precise annotations, such as training machine learning models. This reliability can be improved through careful visual design of campaigns and statistical analysis after data collection to address inconsistencies. Launched in 2019, the ForestEyes project combines citizen science, image analysis, and machine learning to detect deforestation in tropical rainforests. Continuous improvements remain necessary to ensure more reliable data. These enhancements may include using different remote sensing images (e.g., Landsat-8 and Sentinel-2), modifying the visual layout (e.g., grid mode and flip view), and adding viewing options such as zoom. This study presents four consecutive ForestEyes campaigns, each designed with targeted enhancements to improve volunteer classification quality. The first three campaigns followed a continuous improvement strategy, with each iteration addressing limitations identified previously. This process led to the use of higher resolution Sentinel-2 imagery in the third campaign, resulting in a notable accuracy increase compared to Landsat-8 data. To further investigate the gains from this improvement, a fourth campaign tested a new “transfer segmentation” approach, which applies segmentation to Sentinel-2 images and transfers the results to Landsat-8 imagery. Volunteer classification accuracy rose from 57.77% in the first campaign (Landsat-8) to 93.33% in the third campaign (Sentinel-2). These findings highlight the value of citizen science in detecting deforestation and underscore the importance of iterative campaign design and collaborative efforts to protect tropical forests.


  • Open Access

    Do Superpixel Segmentation Methods Influence Deforestation Image Classification?

    by Hugo Resende, Fabio A. Faria, Eduardo B. Neto, Isabela Borlido, Victor Sundermann, Silvio Jamil F. Guimarães, Álvaro L. Fazenda

    Image segmentation is a crucial step in various visual applications, including environmental monitoring through remote sensing. In the context of the ForestEyes project, which combines citizen science and machine learning to detect deforestation in tropical forests, image segments are used for labeling by volunteers and subsequent model training. Traditionally, the Simple Linear Iterative Clustering (SLIC) algorithm is adopted as the segmentation method. However, recent studies have indicated that other superpixel-based methods outperform SLIC in remote sensing image segmentation, and might suggest that they are more suitable for the task of detecting deforested areas. In this sense, this study investigated the impact of the four best segmentation methods, together with SLIC, on the training of classifiers for the target application. Initially, the results showed little variation in performance among segmentation methods, even when selecting the top five classifiers using the PyCaret AutoML library. However, by applying a classifier fusion approach (ensemble of classifiers), noticeable improvements in balanced accuracy were observed, highlighting the importance of both the choice of segmentation method and the combination of machine learning-based models for deforestation detection tasks.

  • Optimizing 2D Packing Strategies for Autoclave Loading Using Deep Reinforcement Learning

    by Victor U. Pugliese, Diogo S. Carvalho, Oseias F. Ferreira, Fabio A. Faria, Francisco S. Melo