Search

It is undeniable that aerial/satellite images can provide useful information for a large variety of tasks. But, since these images are always taken from above, some applications can benefit from complementary information provided by other perspective views of the scene, such as ground-level images. Despite a large number of public repositories for both georeferenced photographs and aerial images, there is a lack of benchmark datasets that allow the development of approaches that exploit the benefits and complementarity of aerial/ground imagery. In this article, we present two new publicly available datasets named AiRound and CV-BrCT. The first one contains triplets of images from the same geographic coordinate with different perspectives of view extracted from various places around the world. Each triplet is composed of an aerial RGB image, a ground-level perspective image, and a Sentinel-2 sample. The second dataset contains pairs of aerial and street-level images extracted from southeast Brazil. We design an extensive set of experiments concerning multiview scene classification, using early and late fusion. Such experiments were conducted to show that image classification can be enhanced using multiview data.

In this work we present BrazilDAM, a novel public dataset based on Sentinel-2 and Landsat-8 satellite images covering all tailings dams cataloged by the Brazilian National Mining Agency (ANM). The dataset was built using georeferenced images from 769 dams, recorded between 2016 and 2019. The time series were processed in order to produce cloud free images. The dams contain mining waste from different ore categories and have highly varying shapes, areas and volumes, making BrazilDAM particularly interesting and challenging to be used in machine learning benchmarks. The original catalog contains, besides the dam coordinates, information about: the main ore, constructive method, risk category, and associated potential damage. To evaluate BrazilDAM’s predictive potential we performed classification essays using state-of-the-art deep Convolutional Neural Network (CNNs). In the experiments, we achieved an average classification accuracy of 94.11% in tailing dam binary classification task. In addition, others four setups of experiments were made using the complementary information from the original catalog, exhaustively exploiting the capacity of the proposed dataset.

Digitization techniques for biomedical images yield disparate visual patterns in radiological exams. These pattern differences, which can be viewed as a domain-shift problem, may hamper the use of data-driven approaches for inference over these images, such as Deep Neural Networks. Another noticeable difficulty in this field is the lack of labeled data, even though in many cases there is an abundance of unlabeled data available. Therefore, an important step in improving the generalization capabilities of these methods and mitigate domain-shift effects is to perform unsupervised or semi-supervised adaptation between different domains of biomedical images. In this work, we propose a novel approach for segmentation of biomedical images based on Generative Adversarial Networks. The proposed method, named Conditional Domain Adaptation Generative Adversarial Network (CoDAGAN), merges unsupervised networks with supervised deep semantic segmentation architectures in order to create a semi-supervised method capable of learning from both unlabeled and labeled data, whenever labeling is available. We conducted experiments to compare our method with traditional and state-of-the-art baselines by using several domains, datasets, and segmentation tasks. The proposed method yielded consistently better results than the baselines in scarce labeled data scenarios, achieving Jaccard values greater than 0.9 and good segmentation quality in most tasks. Unsupervised Domain Adaptation results were observed to be close to the Fully Supervised Domain Adaptation used in the traditional procedure of fine-tuning pretrained networks.

This paper introduces a two-step hyper- and multi-spectral image classification approach. The first step relies on the use of a genetic programming (GP) framework to both select and combine appropriate bands. The second step is concerned with the image classification itself. We present two strategies for multi-class classification problems based on the combination of GP-based indices defined in binary classification scenarios. Performed experiments involving well-known and widely-used datasets demonstrate that the proposed approach yields comparable or better effectiveness performance when compared to several traditional baselines.

Remote Sensing Images (RSI) have been used as a major source of data, particularly with respect to the creation of thematic maps. This process is usually modeled as a supervised learning task where the system needs to learn the patterns of interest provided by the user and assign a class to the rest of the image regions. Thus, it is common to have images obtained from different sensors, which could improve the quality of thematic maps. However, this requires the creation of techniques to properly encode and combine the different properties of the images. So, this paper proposes a boosting-based technique for classification of regions in RSI that manages to encode features extracted from different sources of data, spectral and spatial domains. The approach is evaluated in an urban and a coffee crop recognition scenarios, achieving statistically better results in comparison with the baselines in urban classification and better results at some baselines for the coffee crop recognition.

Sistemas de Informação Geográfica (SIGs) são ferramentas de computador que analisam, armazenam, manipulam e visualizam informações geográficas em mapas. Permitindo às pessoas mais facilmente verem, interpretarem e entenderem os padrões e relacionamentos, os SIGs são importantes ferramentas que ajudam a tomar decisões em diversas questões relacionadas às políticas públicas de monitoramento ambiental e urbano. Além disso, são instrumentos essenciais para o planejamento do desenvolvimentoeconômico e da análise das atividades de agricultura e extração mineral. Nesse contexto, Imagens de Sensoriamento Remoto (ISRs) têm sido usadas como uma das mais importantes fonte de dados, principalmente com relação à criação de mapas temáticos. Esse processo é geralmente modelado como um problema de classificaçãosupervisionada em que o sistema precisa aprender os padrões de interesse fornecidos pelo usuário e atribuir uma classe ao restante das regiões da imagem. Associados à natureza das ISRs, existem vários desafios que podem ser ressaltados: (1) são imagens georreferenciadas, ou seja, existe uma coordenada geográfica associada à cada pixel da imagem; (2) os dados comumente captam frequências específicas em todo o espectro eletromagnético em vez de apenas o espectro visível, o que exige o desenvolvimento de algoritmos específicos para descrever os padrões; (3) o nível de detalhe de cadaimagem pode variar, resultando em imagens com diferentes resoluções espacial e de pixel, porém cobrindo a mesma área de interesse; (4) devido às imagens de alta resolução de pixel, algoritmos de processamento eficientes são desejáveis. Assim, é muito frequente ter imagens obtidas a partir de diferentes sensores, o que poderia melhorar a qualidade dos mapas temáticos gerados. No entanto, isso exige a criação de técnicas capazes de codificar e combinar adequadamente as diferentes propriedades das imagens. Desse modo, esta dissertação propõe duas novas técnicas de classificação deregiões em ISR, com a habilidade de codificar características extraídas de diferentes sensores, dos domínios espacial e espectral. O principal objetivo é o desenvolvimento de um arcabouço capaz de explorar a diversidade entre esses diferentes tipos de dados para alcançar alta acurácia na criação de mapas temáticos. A primeira abordagem proxiposta, chamada de Voto Majoritário Dinâmico, combina dados de diferentes domínios usando uma estratégia de fusão de dados no nível de decisão que explora a especialidade de diferentes classificadores e realiza a combinação deles para a decisão final de cada pixel na construção do mapa temático. Essa abordagem foi criada para lidarapenas com problemas multi-classe e foi avaliada em uma base de dados de cobertura de um cenário urbano. A segunda abordagem proposta realiza a combinação de dados nos três níveis fusão: pixel, características e decisão, assim explorando diferentes tiposde características, extraídas de diferentes sensores usando uma estratégia baseada em boosting de classificadores. Diferentemente do Voto Majoritário Dinâmico, esse método foi projetado para lidar com problemas binários e multi classe, e avaliado na mesma base de dados multi classe em cenário urbano, mas também no problema binário parareconhecimento de plantações de café. Geographic Information Systems (GIS) are computer-based tools that analyzes, stores, manipulates and visualizes geographic information on maps. Enabling people to more easily see, interpret, and understand patterns and relationships, they are important tools, which helps to make decisions on diverse issues related to public policies of urban and environmental monitoring. Furthermore, they are essential tools for economic development planning and analysis in activities of agriculture and mineral extraction. In this context, Remote Sensing Images (RSIs) have been used as a major source of data, particularly with respect to the creation of thematic maps. This process is usually modeled as a supervised classification problem where the system needs to learn the patterns of interest provided by the user and assign a class to the rest of the image regions. Associated with the nature of RSIs, there are several challenges that can be highlighted: (1) they are georeferenced images, i.e., a geographic coordinate is associated for each pixel; (2) the data commonly captures specific frequencies across the electromagnetic spectrum instead of the visible spectrum, which requires the development of specific algorithms to describe patterns; (3) the detail level of each data may vary, resulting in images with different spatial and pixel resolution, but covering the same area; (4) due to the high pixel resolution images, efficient processing algorithms are desirables. Thus, it is very common to have images obtained from different sensors, which could improve the quality of thematic maps generated. However, this requires the creation of techniques to properly encode and combine the different properties of the images. Thus, this dissertation proposes two new techniques for classification of regions in remote sensing images that manage to encode features extracted from different sources of data, spectral and spatial domains. The major objective is the development of a framework able to exploit the diversity of these different types of features to achieve high degrees of accuracy in the creation of thematic maps. The first approach proposed, as named Dynamic Majority Vote, combine the data from different domains using a strategy of fusion at the decision level, which exploits the specialty of different classifiers and combine them for a final decision for each pixel in the thematic map. This approach was created to handle only with multiclass problems and was evaluated in a urban land cover scenario dataset. The second approach proposed, is a combination of the pixel, feature and decision levels fusion, which exploit different types of features, extracted from various sensors using an boosting-based approach. Differently of the Dynamic Majority Vote, this method can handle with binary and multiclass problems, and was evaluated in the same urban land cover multiclass scenario dataset, and also in a coffee crop recognition binary problem.

With the advances on science, a powerful computational infrastructure is desirable to increase the performance of experiments. The same holds true for research on the Computer Vision field, which deals with large amounts of data and also requires intensive computation. Nevertheless, the administration of a computational infrastructure involves many tasks, such as system configuration, preventive maintenance and storage management, which becomes very challenging for many research groups. With that in mind, this work proposes an infrastructure model to assist researchers with focus on Computer Vision. We conducted a series of tests to evaluate the performance of our model and the processing power of our infrastructure.

Foodborne diseases are a growing public health problem worldwide with Campylobacter and Salmonella being the most common and widely distributed causative agents. These Gramnegative bacteria are common inhabitant of the gut of warm-blooded animals, especially livestock, being transmitted to humans primarily through the consumption of contaminated food of animal origin. Poultry meat and derivatives are regarded as the most common source of human salmonellosis and campylobacteriosis.