Skin cancer is by far the most common type of cancer. Early detection is the key to increase the chances for successful treatment significantly. Currently, Deep Neural Networks are the state-of-the-art results on automated skin cancer classification. To push the results further, we need to address the lack of annotated data, which is expensive and require much effort from specialists. To bypass this problem, we propose using Generative Adversarial Networks for generating realistic synthetic skin lesion images. To the best of our knowledge, our results are the first to show visually-appealing synthetic images that comprise clinically-meaningful information.

Deep learning models show remarkable results in automated skin lesion analysis. However, these models demand considerable amounts of data, while the availability of annotated skin lesion images is often limited. Data augmentation can expand the training dataset by transforming input images. In this work, we investigate the impact of 13 data augmentation scenarios for melanoma classification trained on three CNNs (Inception-v4, ResNet, and DenseNet). Scenarios include traditional color and geometric transforms, and more unusual augmentations such as elastic transforms, random erasing and a novel augmentation that mixes different lesions. We also explore the use of data augmentation at test-time and the impact of data augmentation on various dataset sizes. Our results confirm the importance of data augmentation in both training and testing and show that it can lead to more performance gains than obtaining new images. The best scenario results in an AUC of 0.882 for melanoma classification without using external data, outperforming the top-ranked submission (0.874) for the ISIC Challenge 2017, which was trained with additional data.

Convolutional neural networks (CNNs) deliver exceptional results for computer vision, including medical image analysis. With the growing number of available architectures, picking one over another is far from obvious. Existing art suggests that, when performing transfer learning, the performance of CNN architectures on ImageNet correlates strongly with their performance on target tasks. We evaluate that claim for melanoma classification, over 9 CNNs architectures, in 5 sets of splits created on the ISIC Challenge 2017 dataset, and 3 repeated measures, resulting in 135 models. The correlations we found were, to begin with, much smaller than those reported by existing art, and disappeared altogether when we considered only the top-performing networks: uncontrolled nuisances (i.e., splits and randomness) overcome any of the analyzed factors. Whenever possible, the best approach for melanoma classification is still to create ensembles of multiple models. We compared two choices for selecting which models to ensemble: picking them at random (among a pool of highquality ones) vs. using the validation set to determine which ones to pick first. For small ensembles, we found a slight advantage on the second approach but found that random choice was also competitive. Although our aim in this paper was not to maximize performance, we easily reached AUCs comparable to the first place on the ISIC Challenge 2017.

Insulators of the electrical power grid are usually installed outdoors, so they suffer from environmental stresses, such as the presence of contamination. Contamination can increase surface conductivity, which can lead to system failures, reducing the reliability of the network. The identification of insulators that have their properties compromised is important so that there are no discharges through its insulating body. To perform the classification of contaminated insulators, this paper presents computer vision techniques for the extraction of contamination characteristics, and a neural network (NN) model for the classification of this condition. Specifically, the Sobel edge detector, Canny edge detection, binarization with threshold, adaptive binarization with threshold, threshold with Otsu and Riddler–Calvard techniques will be evaluated. The results show that it is possible to have an accuracy of up to 97.50% for the classification of contaminated insulators from the extraction of characteristics with computer vision using the NN for the classification. The proposed model is more accurate than well‐established models such as support‐vector machine (SVM), k‐nearest neighbor (k‐NN), and ensemble learning methods. This showed that optimizing the model's parameters can make it superior to solve the problem in question.

Contamination on insulators may increase the surface conductivity of the insulator, and as a consequence, electrical discharges occur more frequently, which can lead to interruptions in a power supply. To maintain reliability in an electrical distribution power system, components that have lost their insulating properties must be replaced. Identifying the components that need maintenance is a difficult task as there are several levels of contamination that are hard to notice during inspections. To improve the quality of inspections, this paper proposes using k-nearest neighbors (k-NN) to classify the levels of insulator contamination based on images of insulators at various levels of contamination simulated in the laboratory. Computer vision features such as mean, variance, asymmetry, kurtosis, energy, and entropy are used for training the k-NN. To assess the robustness of the proposed approach, a statistical analysis and a comparative assessment with well-consolidated algorithms such as decision tree, ensemble subspace, and support vector machine models are presented. The k-NN showed up to 85.17% accuracy using the k-fold cross-validation method, with an average accuracy higher than 82% for the multi-classification of contamination of insulators, being superior to the compared models.

Due to the increasing number of violence cases, there is a high demand for efficient monitoring systems, however, these systems can be susceptible to failure. Therefore, this work proposes the analysis and application of low-cost Convolutional Neural Networks (CNNs) techniques to automatically recognize and classify suspicious events. Thus, it is possible to alert and assist the monitoring process with a reduced deployment cost. For this purpose, a dataset with violence and non-violence actions in scenes of crowded and non-crowded environments was assembled. The mobile CNNs architectures were adapted and obtained a classification accuracy of up to 92.05%, with a low number of parameters. To demonstrate the models' validity, a prototype was developed by using an embedded Raspberry Pi platform, able to execute a model in real-time with 4 frames-per-second of speed. In addition, a warning system was developed to recognize pre-fight behavior and anticipate violent acts, alerting security to potential situations.

No abstract

No abstract