Aerial scene classification in remote sensing presents a significant challenge due to high intra-class variability and the different scales and orientations of the objects within dataset images. While deep learning architectures are commonly used for scene classification tasks and in the remote sensing area, Deep metric learning (DML) offers a more adaptable solution for more challenging classification scenarios by learning the characteristics of each class. This study exploits the usage of DML approaches for aerial scene classification tasks, analyzing their behavior with different pre-trained Convolutional Neural Networks (CNNs), and their combination through evolutionary computation algorithms. Our experiments show that DML approaches can achieve better classification results as compared to traditional pre-trained CNNs for three well-known remote sensing aerial scene datasets. Furthermore, we found that using the Univariate Marginal Distribution Algorithm (UMDA) to construct the final ensemble of varied DML-based classifiers is essential for achieving consistency and high-accuracy results across all datasets, improving the state-of-the-art by over 5.6%.

The conservation of tropical forests is a topic of significant social and ecological relevance due to their crucial role in the global ecosystem. Unfortunately, deforestation and degradation impact millions of hectares annually, necessitating government or private initiatives for effective forest monitoring. This study introduces a novel framework that employs the Univariate Marginal Distribution Algorithm (UMDA) to select spectral bands from Landsat-8 satellite optical sensor, optimizing the representation of deforested areas. This selection guides a semantic segmentation architecture, DeepLabv3+, enhancing its performance. Experimental results revealed several band compositions that achieved superior balanced accuracy compared to commonly adopted combinations for deforestation detection, utilizing segment classification via a Support Vector Machine (SVM). Moreover, the optimal band compositions identified by the UMDA-based approach improved the performance of the DeepLabv3+ architecture, surpassing state-of-the-art approaches compared in this study. The observation that a few selected bands outperform the total contradicts the data-driven paradigm prevalent in the deep learning field. Therefore, this suggests an exception to the conventional wisdom that 'more is always better'.

The conservation of tropical forests is a topic of significant social and ecological relevance due to their crucial role in the global ecosystem. Unfortunately, deforestation and degradation impact millions of hectares annually, requiring government or private initiatives for effective forest monitoring. However, identifying deforested regions in satellite images is challenging due to data imbalance, image resolution, low-contrast regions, and occlusion. Superpixel segmentation can overcome these drawbacks, reducing workload and preserving important image boundaries. However, most works for remote-sensing images do not exploit recent superpixel methods. In this work, we evaluate 16 superpixel methods in satellite images to support a deforestation detection system in tropical forests. We also assess the performance of superpixel methods for the target task, establishing a relationship with segmentation methodological evaluation. According to our results, ERS, GMMSP, and DISF perform best on undersegmentation error (UE), boundary recall (BR), and similarity between image and reconstruction from superpixels (SIRSs), respectively, whereas ERS has the best tradeoff with compactness index (CO) and Reg. In classification, SH, DISF, and ISF perform best on RGB, UMDA, and PCA compositions, respectively. According to our experiments, superpixel methods with better tradeoffs among delineation, homogeneity, compactness, and regularity are more suitable for identifying good superpixels for deforestation detection tasks.

Deep learning architectures have achieved promising results in different areas (e.g., medicine, agriculture, and security). However, using those powerful techniques in many real applications becomes challenging due to the large labeled collections required during training. Several works have pursued solutions to overcome it by proposing strategies that can learn more for less, e.g., weakly and semi-supervised learning approaches. As these approaches do not usually address memorization and sensitivity to adversarial examples, this paper presents three deep metric learning approaches combined with Mixup for incomplete-supervision scenarios. We show that some state-of-the-art approaches in metric learning might not work well in such scenarios. Moreover, the proposed approaches outperform most of them in different datasets.

Originally, the ForestEyes project aims to detect deforestation in tropical forests based on citizen science (CS) and machine learning (ML) approaches, in which the volunteers analyze and label segments of remote sensing images to build new training sets for creating different classification models. In previous work, only three modules related to CS have been proposed. In this letter, two new modules are created: 1) organization and selection and 2) ML. Therefore, these modules turn the ForestEyes project a more robust system in the deforestation detection task, building high-confidence labeled collections, increasing the monitoring coverage, and decreasing volunteer dependence. Performed experiments show that volunteers create better data sets than those based on automatic PRODES-based approaches, selecting the most relevant samples and discarding noisy segments that might disrupt ML techniques. Finally, the results showed the feasibility of allying CS with ML for rainforest deforestation detection task.

Convolutional Neural Networks (CNN) have been being widely employed to solve the challenging remote sensing task of aerial scene classification. Nevertheless, it is not straightforward to find single CNN models that can solve all aerial scene classification tasks, allowing the development of a better alternative, which is to fuse CNN-based classifiers into an ensemble. However, an appropriate choice of the classifiers that will belong to the ensemble is a critical factor, as it is unfeasible to employ all the possible classifiers in the literature. Therefore, this work proposes a novel framework based on meta-heuristic optimization for creating optimized ensembles in the context of aerial scene classification. The experimental results were performed across nine meta-heuristic algorithms and three aerial scene literature datasets, being compared in terms of effectiveness (accuracy), efficiency (execution time), and behavioral performance in different scenarios. Our results suggest that the Univariate Marginal Distribution Algorithm shows more effective and efficient results than other commonly used meta-heuristic algorithms, such as Genetic Programming and Particle Swarm Optimization.

The aerial scene-classification task is a challenging problem to remote sensing area with important applicability to civil and military affairs. A technique that has achieved excellent results in this task is the convolutional neural network (CNN). CNNs are powerful semantic-level feature-extraction techniques successfully applied to many application domains. Nevertheless, many works in the literature have shown that a single CNN cannot solve all kinds of application domains properly. Hence, an alternative solution might be the joining of CNN architectures as an ensemble of classifiers. In this sense, this letter proposes a new strategy of deep feature-based classifier fusion through a meta-feature engineering approach based on the Kaizen programming (KP) technique for the aerial scene-classification task. KP is a technique that continuously improves partial solutions and combines them into a complete solution. In the context, a partial solution is a meta-feature, and a complete solution is an ensemble of classifiers. In our experiments on three different public data sets, we show that KP can automatically engineer meta-features that significantly improve the performance of a stacked classifier while reducing the number of total meta-features.

Scientific projects involving volunteers for analyzing, collecting data, and using their computational resources, known as Citizen Science (CS), have become popular due to advances in information and communication technology (ICT). Many CS projects have been proposed to involve citizens in different knowledge domain such as astronomy, chemistry, mathematics, and physics. This work presents a CS project called ForestEyes, which proposes to track deforestation in rainforests by asking volunteers to analyze and classify remote sensing images. These manually classified data are used as input for training a pattern classifier that will be used to label new remote sensing images. ForestEyes project was created on the Zooniverse.org CS platform, and to attest the quality of the volunteers' answers, were performed early campaigns with remote sensing images from Brazilian Legal Amazon (BLA). The results were processed and compared to an oracle classification (PRODES - Amazon Deforestation Monitoring Project). Two and a half weeks after launch, more than 35,000 answers from 383 volunteers (117 anonymous and 266 registered users) were received, completing all 2050 tasks. The ForestEyes campaigns' results have shown that volunteers achieved excellent effectiveness results in remote sensing image classification task. Furthermore, these results show that CS might be a powerful tool to quickly obtain a large amount of high-quality labeled data.

Many works in the literature have used machine learning techniques to solve their classification problems in different knowledge areas, e.g., medicine, agriculture, and remote sensing. Since there is no a single machine learning technique that achieves the best results for all kind of applications, a good alternative is the fusion of classification techniques, also known as multiple classifier systems (MCS). A common challenge in MCS is the selection of a few classifiers among many classifiers that are available in the literature; using all possible classifiers is not a feasible alternative. The choice of the classifiers becomes an essential factor, i.e., we need an ensemble selection approach. In this work, we propose a novel graph-based approach for static ensemble selection (GASES) to find or choose the best classifier set for remote sensing image classification. Experiments demonstrate that GASES improves performance by up to 70% over different baseline approaches when fusing classifiers. It decreases the number of classifiers used while retaining the effectiveness of using all of the classifiers. Furthermore, our proposed method is a more straightforward and intuitive technique for static ensemble selection scheme than other baseline approaches such as Consensus and Kendall.

Fruit flies are of huge biological and economic importance for the farming of different countries in the World, especially for Brazil. Brazil is the third largest fruit producer in the world with 44 million tons in 2016. The direct and indirect losses caused by fruit flies can exceed USD 2 billion, putting these pests as one of the biggest problems of the world agriculture. In Brazil, it is estimated that the economic losses directly related to production, the cost of pest control and in the loss of export markets, are between USD 120 and 200 million/year. The species of the genus Anastrepha are among the fruit flies economically important in the America tropics and subtropics with approximately 300 known species, of which 120 are recorded in Brazil. However, few species are economically important in Brazil and are considered pests of quarantine significance by regulatory agencies. In this sense, the development of automatic and semi-automatic tools for fruit fly species identification of the genus Anastrepha can assist the few existing specialists to reduce the insect analysis time and the economic losses related to these agricultural pests. We propose to apply mid-level image representations based on local descriptors for fruit fly identification tasks of three species of the genus Anastrepha. In our experiments, several local image descriptors based on keypoints and machine learning techniques have been studied for the target task. Furthermore, the proposed approaches have achieved excellent effectiveness results when compared with a state-of-art technique.

Brain connectivity analysis is a critical component of ongoing human connectome projects to decipher the healthy and diseased brain. Recent work has highlighted the power-law (multi-time scale) properties of brain signals; however, there remains a lack of methods to specifically quantify short- vs. long- time range brain connections. In this paper, using detrended partial cross-correlation analysis (DPCCA), we propose a novel functional connectivity measure to delineate brain interactions at multiple time scales, while controlling for covariates. We use a rich simulated fMRI dataset to validate the proposed method, and apply it to a real fMRI dataset in a cocaine dependence prediction task. We show that, compared to extant methods, the DPCCA-based approach not only distinguishes short and long memory functional connectivity but also improves feature extraction and enhances classification accuracy. Together, this paper contributes broadly to new computational methodologies in understanding neural information processing.

Cocaine dependence devastates millions of human lives. Despite of a variety of treatments, there is a very high rate of individual relapse to drug use. In the last decade, functional magnetic resonance imaging (fMRI) proved to be a powerful tool to diagnose and understand different pathologies. This work provides advances in the identification of cocaine dependence and in the relapse prediction based on fMRI classification. We improve the traditional methodology of the literature called multi-voxel pattern analysis (MVPA), which is used for feature extraction and classification. In addition, we propose new features that use specific functional connectivity measures. An extensive evaluation was conducted comparing our methodology with MVPA, as well as, several learning methods with distinct feature sets. We could identify the neural patterns that lead to improve classification accuracies and evaluate the advantages of employing an information fusion approach through an ensemble of classifiers. Experimental results show an improvement of final accuracy over the state-of-the-art methods.

In this paper, we explore transformed spaces, represented by image illuminant maps, to propose a methodology for selecting complementary forms of characterizing visual properties for an effective and automated detection of image forgeries. We combine statistical telltales provided by different image descriptors that explore color, shape, and texture features. We focus on detecting image forgeries containing people and present a method for locating the forgery, specifically, the face of a person in an image. Experiments performed on three different open-access data sets show the potential of the proposed method for pinpointing image forgeries containing people. In the two first data sets (DSO-1 and DSI-1), the proposed method achieved a classification accuracy of 94% and 84%, respectively, a remarkable improvement when compared with the state-of-the-art methods. Finally, when evaluating the third data set comprising questioned images downloaded from the Internet, we also present a detailed analysis of target images.

Phenology is the study of periodic natural phenomena and their relationship to climate. Usually, phenology studies consider the identification of patterns on temporal data. In those studies, several phenological change patterns are often encoded in time series for analysis and knowledge extraction. In this paper, we evaluate the effectiveness of several time series similarity functions in the task of classifying time series related to phonological phenomena characterized by near-surface vegetation indices extracted from images. In addition, we performed a correlation analysis to identify potential candidates for combination.

Coffee crop recognition in remote sensing images is a complex task. It poses several challenges due to different spectral responses and texture patterns that can be extracted from coffee regions. This paper presents a novel framework for combining different classifiers using support vector machine technique (SVM), which try to learn with each one of classifiers previews experiences (meta-learning). We investigate the combination of seven learning methods and seven image descriptors aiming at creating low-cost classifiers for coffee crops recognition. The objective is to provide an effective mechanism for coffee crop recognition by fusion of region-based classifiers in remote sensing images. The experiments showed that the proposed framework for fusion of classifiers produces better results than the traditional majority voting fusion approach and all base classifiers tested.

This work presents a new approach for automatic recognition of coffee crops in RSIs. The method applies an approach based on Genetic Programming (GP) to combine texture and spectral information encoded by image descriptors. Experiments show that the proposed method yields slightly better results than the traditional MaxVer approach.