No abstract

The use of photovoltaic systems for clean electrical energy has increased. However, due to their low efficiency, researchers have looked for ways to increase their effectiveness and improve their efficiency. The Maximum Power Point Tracking (MPPT) inverters allow us to maximize the extraction of as much energy as possible from PV panels, and they require algorithms to extract the Maximum Power Point (MPP). Several intelligent algorithms show acceptable performance; however, few consider using Artificial Neural Networks (ANN). These have the advantage of giving a fast and accurate tracking of the MPP. The controller effectiveness depends on the algorithm used in the hidden layer and how well the neural network has been trained. Articles over the last six years were studied. A review of different papers, reports, and other documents using ANN for MPPT control is presented. The algorithms are based on ANN or in a hybrid combination with FL or a metaheuristic algorithm. ANN MPPT algorithms deliver an average performance of 98% in uniform conditions, exhibit a faster convergence speed, and have fewer oscillations around the MPP, according to this research.

Automatic sign language recognition is a challenging task in machine learning and computer vision. Most works have focused on recognizing sign language using hand gestures only. However, body motion and facial gestures play an essential role in sign language interaction. Taking this into account, we introduce an automatic sign language recognition system based on multiple gestures, including hands, body, and face. We used a depth camera (OAK-D) to obtain the 3D coordinates of the motions and recurrent neural networks for classification. We compare multiple model architectures based on recurrent networks such as Long Short-Term Memories (LSTM) and Gated Recurrent Units (GRU) and develop a noise-robust approach. For this work, we collected a dataset of 3000 samples from 30 different signs of the Mexican Sign Language (MSL) containing features coordinates from the face, body, and hands in 3D spatial coordinates. After extensive evaluation and ablation studies, our best model obtained an accuracy of 97% on clean test data and 90% on highly noisy data.

No abstract

The measures most commonly used in current literature to compute the roundness of digital objects are derivations of the form factor based on area and perimeter computations. However, these measures are highly dependent on image resolution and sensitive to shape variations. In this article, a new measure is proposed. This measure takes into consideration the dominant geometry of objects, avoiding the use of such parameters as area, perimeter and Ferret's diameter. The proposed measure is easy to compute, and since it is a distribution of probability based on the radius, it is invariant to abrupt changes in contours or to shape resolution. In order to show the performance of this measure, it is compared with three other recently proposed measures: factor shape, which is recommended by the American Standard Test Measurement, mean roundness and radius ratio.