Sliding Window-Based Deep Learning Approach for Solar Power Forecasting in Malaysia Utility-Scale PV Systems

The concept of Scene-to-Speech (STS) is to recognize elements in a captured image or a video clip to speak loudly an informative textual content that describes the scene. The contemporary progression in convolution neural network (CNN) allows us to attain object recognition procedures, in real-time, on mobile handled devices. Considerable number of applications has been developed to perform object recognition in scenes and say loudly their relevant descriptive messages. However, the employment of multiple trained deep learning (DL) models is not fully supported. In our previous work, a mobile application that can capture images and can recognize the objects contained in them was developed. It constructs descriptive sentences and speak them in Arabic and English languages. The notion of employing multi-trained DL models was used but no experimentation was conducted. In this article, we extend our previous work to perform required assessments while using multiple trained DL models. The main aim is to show that the deployment of multiple models approach can reduce the complexity of having one large compound model, and can enhance the prediction time. For this reason, we examine the prediction accuracy for single DL model-based recognition and multiple DL model-based recognition scenarios. The assessments results showed significant improvement in the prediction accuracy and in the prediction time. In the other hand, from the end user aspect, the application is designed primarily for visually impaired people to assist them in understanding their surroundings. In this context, we conduct a usability study to evaluate the usability of the proposed application with normal people and with visually impaired people. In fact, participants showed large interest in using the mobile application daily.

Advancing short-term solar irradiance forecasting accuracy through a hybrid deep learning approach with Bayesian optimization

MLSPred-bench: Transforming electroencephalography (EEG) datasets into machine learning-ready epileptic seizure prediction benchmarks

PurposeAccurate assessment of the percentage of total body surface area (%TBSA) burned is crucial in managing burn injuries. It is difficult to estimate the size of an irregular shape by inspection. Many articles reported the discrepancy of estimating %TBSA burned by different doctors. We set up a system with multiple deep learning (DL) models for %TBSA estimation, as well as the segmentation of possibly poor-perfused deep burn regions from the entire wound. MethodsWe proposed boundary-based labeling for datasets of total burn wound and palm, whereas region-based labeling for the dataset of deep burn wound. Several powerful DL models (U-Net, PSPNet, DeeplabV3+, Mask R-CNN) with encoders ResNet101 had been trained and tested from the above datasets. With the subject distances, the %TBSA burned could be calculated by the segmentation of total burn wound area with respect to the palm size. The percentage of deep burn area could be obtained from the segmentation of deep burn area from the entire wound. ResultsA total of 4991 images of early burn wounds and 1050 images of palms were boundary-based labeled. 1565 out of 4994 images with deep burn were preprocessed with superpixel segmentation into small regions before labeling. DeeplabV3+ had slightly better performance in three tasks with precision: 0.90767, recall: 0.90065 for total burn wound segmentation; precision: 0.98987, recall: 0.99036 for palm segmentation; and precision: 0.90152, recall: 0.90219 for deep burn segmentation. ConclusionCombining the segmentation results and clinical data, %TBSA burned, the volume of fluid for resuscitation, and the percentage of deep burn area can be automatically diagnosed by DL models with a pixel-to-pixel method. Artificial intelligence provides consistent, accurate and rapid assessments of burn wounds.

This paper presents a method of multiple Deep Learning model-based visibility estimation for a fixed and drone camera under adversarial conditions. High visibility is required for safe flights at airports, cities, and fields. On the other hand, local weather conditions can suddenly vary during the flight. Fog and clouds are known to cause invisibility that impedes critical issues for flights. Due to expensive systems, visibility measurement sensors have been used at very limited airports. For this, lower-cost methods with portable apparatus have been desired for flight routines. Therefore, this paper proposes a camera-based visibility estimation method using multiple Deep Learning models. However, even state-of-the-art models can be degraded by strong illumination like sunbeams and artificial lighting. For this issue, DeepReject is introduced to exclude input images. DeepScene is proposed to apply for panoptic segmentation of objects. In DeepDist, fog levels can be estimated by apparent changes in each geo-tagged object, i.e., vehicle, and mountain. An image-data regression-based DeepVis is also proposed to apply for non-geo-tagged scenes, i.e., sky and clouds. Using four DeepXs, experimental results show the robustness, stability, and accuracy of the proposed multiple DL models for the unmanned aerial vehicle flight conditions over a single DL.

In the prioritized vehicle traffic environment, motorized transportation has been obtaining more spatial and economic resources, posing potential threats to the travel quality and life safety of non-motorized transportation participants. It is becoming urgent to improve the safety situation of non-motorized transportation participants. Most previous studies have focused on the psychological aspects of pedestrians and cyclists exposed to the actual road environment rather than quantifying the objective safety hazards, which has led to a non-rigorous evaluation of their basic safety situation. An integrated processing approach is proposed to comprehensively and objectively evaluate the overall safety level of non-motorized transportation participants on each road segment. Our main contributions include (1) the universal approach is established to automatically identify hazard scenarios related to non-motorized transportation and their direct causing factors from street view images based on multiple deep learning models; (2) a seed points spreading algorithm is designed to convert semantic images into target detection results with detail contour, which breaks the functional limitation of these two types of methods to a certain extent; (3) The safety situation of non-motorized transportation on various road sections in Gulou District, Nanjing, China has been evaluated and based on this, a series of suggestions have been put forward to guide the better adaptation among multiple transportation participants.

Sinus X-ray imaging is still used in the initial evaluation of paranasal sinusitis, which is diagnosed by the opacification or air/fluid level in the sinuses and best seen in the Waters' view of the paranasal sinus (PNS). The objective of this study was to investigate the feasibility of recognizing the maxillary sinusitis features using PNS X-ray images, as well as to propose the most effective method of determining a reasonable consensus using multiple deep learning models. A total of 4,860 patients, which included 2,430 normal and maxillary sinusitis subjects each, underwent Waters' view PNS X-ray scan. The datasets were randomly split into training (70%), validation (15%), and test (15%) subsets. We implemented a majority decision algorithm to determine a reasonable consensus using three multiple convolutional neural network (CNN) models: VGG-16, VGG-19, and ResNet-101. The performance of sinusitis detection was evaluated with quantitative accuracy (ACC) and activation maps. We compared the results of our approaches with ACC and activation maps. ACC [and area under the curve (AUC)] of the internal test dataset was evaluated as 87.4% (0.891), 90.8% (0.891), 93.7% (0.937), and 94.1% (0.948) for VGG-16, VGG-19, ResNet-101, and the majority decision, respectively. ACC (and AUC) of the external test dataset was evaluated as 87.58% (0.877), 87.58% (0.877), 92.12% (0.929), and 94.12% (0.942) for VGG-16, VGG-19, ResNet-101, and the majority decision, respectively. Majority decision algorithms can detect missing and correct lesions using a compensation function of the majority decision. The majority decision algorithm showed high accuracy and significantly more accurate lesion detection compared with those of individual CNN models. The proposed deep learning method with PNS X-ray images can be used as an adjunct to classify maxillary sinusitis.

In response to the problem of low forecasting accuracy in wind and solar power outputs, this study proposes a joint forecasting method for wind and solar power outputs by using their spatiotemporal correlation. First, autocorrelation analysis and causal testing are used to screen the forecasting factors. Then, a convolutional neural network–long short-term memory (CNN-LSTM) is constructed and trained to extract features effectively. Finally, the independent, ensemble, and joint forecasting effects are compared, using a certain clean energy base as the research object. Results show that the forecasting accuracy of the ensemble wind and solar power outputs is better than that of independent forecasting. The joint forecasting method can improve the forecasting accuracy of wind power by 20% but slightly affects the forecasting accuracy of solar power.

Deep learning (DL) systems are increasingly used in security-related fields, where the accuracy and predictability of DL systems are critical. However the DL models are difficult to test and existing DL testing relies heavily on manually labeled data and often fails to expose erroneous behavior for corner inputs. In this paper, we propose Differential Combination Testing (DCT), an automated DL testing tool for systematically detecting the erroneous behavior of more corner cases without relying on manually labeled input data or manually checking the correctness of the output behavior. Our tool aims at automatically generating test cases, that is, applying image combination transformations to seed images to systematically generate synthetic images that can achieve high neuron coverage and trigger inconsistencies between multiple similar DL models. In addition, DCT utilizes multiple DL models with similar functions as cross-references, so that input data no longer must be manually marked and the correctness of output behavior can be automatically checked. The results show that DCT can find thousands of erroneous corner behaviors in the most commonly used DL models effectively and quickly, which can better detect the reliability and robustness of DL systems.

Analysis of the impact of irradiance, temperature and tilt angle on the performance of grid-connected solar power plant

The recent rapid uptake of solar photovoltaic system in Malaysia has introduced new challenges to the operation of the power grid, particularly on the voltage profile of distribution networks. An intermittent nature of the voltage fluctuation that caused by solar output remain as important topic as how these rapid voltage changes will affect the operation of the network. Hence, this paper presents a quantification of voltage profiles intermittency for small-scale solar photovoltaic system under different loading conditions. Actual field measurements have been performed for both high and low loading conditions. Mainly two parameters have been analysed for this work, namely voltage profile and solar power output. An intermittency quantity of voltage profile during solar generation for day time is higher than night time, which indicates high intermittent in solar irradiance causing rapid fluctuation in solar generation. Furthermore, increased voltage magnitude was observed during the low load condition. The high voltage intermittency caused by rapid fluctuation of solar output and voltage rise during low load giving a negative effect on distribution network. It is expected that grid-connected battery energy storage system could serve as the potential solution to alleviate both solar intermittency and voltage raise issues.

The interest for solar Photovoltaic (PV) systems installation has been recently increasing due to their reducing cost and high electricity price from the utilities. However, the PV power generation is sensitive to variable solar irradiation, which causes technical difficulties related to power quality and control. Severity of those problems directly depends on PV penetration level, grid design and geographic location. This paper explores the flicker indices of grid-tied PV power plant, one of the major power quality issues. The study has been carried out on a real grid-integrated PV system, situated at Engineering Center, Florida International University (FIU). The field measurements of various solar output power and flicker quantifying indices at the point of common coupling (PCC) have been recorded using data acquisition system and power quality meter respectively. The collected data has been categorized and analyzed for voltage fluctuations and flicker. Different approaches have been done to investigate the influence of PV power plant on the flicker indices. Results show that solar output power has a marginal effect towards flicker indices with existing penetration level. Load connection and variation are the key indicators of high flicker value during observation time.

Fully automated learning and predict price of aquatic products in Taiwan wholesale markets using multiple machine learning and deep learning methods

In this paper, we propose a spatio-temporal predictive network with attention weighting of multiple physical Deep Learning (DL) models for videos with various physical properties. Previous approaches have been models with multiple branches for difference properties in videos, but the outputs of branches have been simply summed even with properties that change in time and space. In addition, it is difficult to train previous models for sufficient representations of physical properties in videos. Therefore, we propose the design of the spatio-temporal prediction network and the training method for videos with multiple physical properties, motivated by the Mixtures of Experts framework. Multiple spatio-temporal DL branches/experts for multiple physical properties and pixel-wise and expert-wise attention mechanism for adaptively integrating outputs of experts, i.e., Spatial-Temporal Gating Networks (STGNs) are proposed. Experts are trained with a vast amount of synthetic image sequences by physical equations and noise models. Instead, the whole network including STGNs is allowed to be trained only with a limited number of real datasets. Experiments on various videos, i.e., traffic, pedestrian, Dynamic Texture videos, and radar images, show the superiority of our proposed approach compared with previous approaches.

Time series data contains temporal ordering, which makes its feature selection different from the normal feature selection. Feature selection in multivariate time series has two tasks: identifying the relevant features and finding their effective window sizes of lagged values. The methods extended from normal feature selection methods do not solve this two-dimensional feature selection problem since they do not take lagged observations of features into consideration. In this paper, we present a method using the Granger causality discovery to identify causal features with effective sliding window sizes in multivariate numerical time series. The proposed method considers the influence of lagged observations of features on the target time series. We compare our proposed feature selection method with several normal feature selection methods on multivariate time series data using three well-known modeling methods. Our method outperforms other methods for predicting future values of target time series. In a real world case study on water quality monitoring data, we show that the features selected by our method contain four out of five features used by domain experts, and prediction performance on our features is better than that on features of domain experts using three modeling methods.