Forest fire detection in aerial vehicle videos using a deep ensemble neural network model

Ensemble deep learning: A review

A Radiogenomic Deep Ensemble Learning Model for Identifying Radionecrosis Following Brain Metastases (BM) Stereotactic Radiosurgery in Patients With Non-small Cell Lung Cancer BM.

Making computer detect desired object have always been an area of interest for humans. Object detection can be implemented using following stages: feature extraction, object localization followed by identifying object in input image. Most of the present-day object detection work is focused around x86 and ARM architectures. Researchers constantly strive to either identify better object detection architectures, updated models, improved model accuracies or reduce prediction time. In this paper, multiple pre-trained Deep Neural Network (DNN) models such as Region Based Convolutional Neural Network (RCNN), Fast RCNN, Faster RCNN. You Only Look Once (YOLO) V3 and Single Shot Multibox Detector (SSD) are used to identify fruits in given input image on RISC- V architecture. In order to bring uniformity across all DNN models, all these models are pre-trained on COCO datasets. Experimental results have shown that out of various DNN models tested for object recognition, YOLO and SSD-MobileNet gives optimum performance in terms of accuracy and inference time on RISC- V architecture.

In the last 50 years, with the growth of cities and increase in the number of vehicles and mobility, traffic has become troublesome. As a result, traffic flow prediction started to attract attention as an important research area. However, despite the extensive literature, traffic flow prediction still remains as an open research problem, specifically for long-term traffic flow prediction. Compared to the models developed for short-term traffic flow prediction, the number of models developed for long-term traffic flow prediction is very few. Based on this shortcoming, in this study, we focus on long-term traffic flow prediction and propose a novel deep ensemble model (DEM). In order to build this ensemble model, first, we developed a convolutional neural network (CNN), a long short-term memory (LSTM) network and a gated recurrent unit (GRU) network as deep learning models, which formed the base learners. In the next step, we combine the output of these models according to their individual forecasting success. We use another deep learning model to determine the success of the individual models. Our proposed model is a flexible ensemble prediction model that can be updated based on traffic data. To evaluate the performance of the proposed model, we use a publicly available dataset. Experimental results show that the developed DEM model has a mean square error of 0.06 and a mean absolute error of 0.15 for single-step prediction; it shows that achieves a mean square error of 0.25 and a mean absolute error of 0.32 for multi-step prediction. We compared our proposed model with many models in different categories; individual deep learning models (i.e., LSTM, CNN, GRU), selected traditional machine learning models (i.e., linear regression, decision tree regression, k-nearest-neighbors regression) and other ensemble models such as random-forest regression. These results also support the claim that ensemble learning models perform better than individual models.

Prediction of wind turbine blade icing fault based on selective deep ensemble model

Forest fires can destroy forest and inflict great damage to the ecosystem. Fortunately, forest fire detection with video has achieved remarkable results in enabling timely and accurate fire warnings. However, the traditional forest fire detection method relies heavily on artificially designed features; CNN-based methods require a large number of parameters. In addition, forest fire detection is easily disturbed by fog. To solve these issues, a lightweight YOLOX-L and defogging algorithm-based forest fire detection method, GXLD, is proposed. GXLD uses the dark channel prior to defog the image to obtain a fog-free image. After the lightweight improvement of YOLOX-L by GhostNet, depth separable convolution, and SENet, we obtain the YOLOX-L-Light and use it to detect the forest fire in the fog-free image. To evaluate the performance of YOLOX-L-Light and GXLD, mean average precision (mAP) was used to evaluate the detection accuracy, and network parameters were used to evaluate the lightweight effect. Experiments on our forest fire dataset show that the number of the parameters of YOLOX-L-Light decreased by 92.6%, and the mAP increased by 1.96%. The mAP of GXLD is 87.47%, which is 2.46% higher than that of YOLOX-L; and the average fps of GXLD is 26.33 when the input image size is 1280 × 720. Even in a foggy environment, the GXLD can detect a forest fire in real time with a high accuracy, target confidence, and target integrity. This research proposes a lightweight forest fire detection method (GXLD) with fog removal. Therefore, GXLD can detect a forest fire with a high accuracy in real time. The proposed GXLD has the advantages of defogging, a high target confidence, and a high target integrity, which makes it more suitable for the development of a modern forest fire video detection system.

With the widespread use of deep neural networks (DNNs) in many areas, more and more studies focus on protecting DNN models from intellectual property (IP) infringement. Many existing methods apply digital watermarking to protect the DNN models. The majority of them either embed a watermark directly into the internal network structure/parameters or insert a zero-bit watermark by fine-tuning a model to be protected with a set of so-called trigger samples. Though these methods work very well, they were designed for individual DNN models, which cannot be directly applied to deep ensemble models (DEMs) that combine multiple DNN models to make the final decision. It motivates us to propose a novel black-box watermarking method in this paper for DEMs, which can be used for verifying the integrity of DEMs. In the proposed method, a certain number of sensitive samples are carefully selected through mimicking real-world DEM attacks and analyzing the prediction results of the sub-models of the non-attacked DEM and the attacked DEM on the carefully crafted dataset. By analyzing the prediction results of the target DEM on these carefully crafted sensitive samples, we are able to verify the integrity of the target DEM. Different from many previous methods, the proposed method does not modify the original DEM to be protected, which indicates that the proposed method is lossless. Experimental results have shown that the DEM integrity can be reliably verified even if only one sub-model was attacked, which has good potential in practice.

Fault occurrences in photovoltaic (PV) modules can hinder the performance of the system, resulting in reduced lifetime and performance of the modules. PV module (PVM) faults if unmonitored can affect the power transmission through the system, thereby creating short circuits that can be hazardous. Unmanned aerial vehicle (UAV)-based monitoring is one of the most common and widely adopted techniques to detect faults in PVM. Visual images of PVM contain the necessary information about the faults, but sometimes, it becomes difficult even for expert professional to work on large amount of image data. Automatic classification of PVM faults using deep learning techniques can help in providing improved analysis and instantaneous results. The present study adopts renowned deep convolution neural network (CNN) models such as MobileNet V2, Inception V3, and Xception for the classification of PVM. The aforementioned models were trained individually, and the classification performances of the models were observed to be 97.03%, 95.55%, and 92.27%, respectively. A hybrid deep ensemble model is proposed in the study that merges all the aforementioned models. The proposed model produced classification accuracy higher than each of the individual model with a value of 99.04%. Automatic classification using deep ensemble model can help in the accurate identification of faults in PVM from images acquired through UAV. Consequently, this computer-aided and quick diagnosis can eliminate the downtime and fire hazards.

Forest fires pose significant threats to forest ecosystems, impacting humans, animals, and plants reliant on these environments. Traditional detection methods rely on handcrafted features like color, motion, and texture, yet achieving accuracy remains challenging. This study introduces a novel approach using a lightweight fire detection method employing Deep Convolution Neural Networks (DCNN), considering temporal aspects for enhanced accuracy. By leveraging DCNN, this study aims to improve forest fire detection capabilities, mitigating the devastating effects of wildfires on both natural habitats and communities. This method represents a promising advancement in the field, offering potential solutions to the ongoing challenge of timely and accurate forest fire detection.

Onychomycosis is a common disease. Emerging noninvasive, real-time techniques such as dermoscopy and deep convolutional neural networks have been proposed for the diagnosis of onychomycosis. However, deep learning application in dermoscopic images has not been reported. To explore the establishment of deep learning-based diagnostic models for onychomycosis in dermoscopy to improve the diagnostic efficiency and accuracy. We evaluated the dermoscopic patterns of onychomycosis diagnosed at Sun Yat-sen Memorial Hospital, Guangzhou, China, from May 2019 to February 2021 and included nail psoriasis and traumatic onychodystrophy as control groups. Based on the dermoscopic images and the characteristic dermoscopic patterns of onychomycosis, we gain the faster region-based convolutional neural networks to distinguish between nail disorder and normal nail, onychomycosis and non-mycological nail disorder (nail psoriasis and traumatic onychodystrophy). The diagnostic performance is compared between deep learning-based diagnosis models and dermatologists. All of 1,155 dermoscopic images were collected, including onychomycosis (603 images), nail psoriasis (221 images), traumatic onychodystrophy (104 images) and normal cases (227 images). Statistical analyses revealed subungual keratosis, distal irregular termination, longitudinal striae, jagged edge, and marble-like turbid area, and cone-shaped keratosis were of high specificity (>82%) for onychomycosis diagnosis. The deep learning-based diagnosis models (ensemble model) showed test accuracy /specificity/ sensitivity /Youden index of (95.7%/98.8%/82.1%/0.809) and (87.5%/93.0%/78.5%/0.715) for nail disorder and onychomycosis. The diagnostic performance for onychomycosis using ensemble model was superior to 54 dermatologists. Our study demonstrated that onychomycosis had distinctive dermoscopic patterns, compared with nail psoriasis and traumatic onychodystrophy. The deep learning-based diagnosis models showed a diagnostic accuracy of onychomycosis, superior to dermatologists.

Automated accurate fire detection system using ensemble pretrained residual network

Unmanned aerial vehicles (UAVs) are increasingly being used in forest fire monitoring and detection thanks to their high mobility and ability to cover areas at different altitudes and locations with relatively lower cost. Traditional fire detection algorithms are mostly based on the RGB color model, but their speed and accuracy need further improvements. This paper proposes a forest fire detection algorithm by exploiting YOLOv3 to UAV-based aerial images. Firstly, a UAV platform for the purpose of forest fire detection is developed. Then according to the available computation power of the onboard hardware, a small-scale of convolution neural network (CNN) is implemented with the help of YOLOv3. The testing results show that the recognition rate of this algorithm is about 83%, and the frame rate of detection can reach more than 3.2 fps. This method has great advantages for real-time forest fire detection application using UAVs.

Background: There is a severe need to detect any kind of fire in a faster and accurate method, especially forest fires to stop huge losses to the human community and the environment losses. The main purpose of the proposal is to identify and evaluate the accuracy of the existing Artificial Intelligence (AI) methods for detecting fire and improve the methods to detect fire in real-world scenarios in faster and accurate methods. Methods: The proposal uses a dataset to train a model, and in addition uses a few test images from an existing database to test the models developed. We develop and test the following neural network models namely, Deep Neural Network (DNN), Recurrent Neural Network (RNN), Convolution Neural Network (CNN), and a combination of RNN and CNN. Findings: A conservative estimate of the yearly losses caused by forest fires in India for the entire nation is 440 crores INR. The loss of biodiversity, soil moisture, nutrients, and other intangible advantages is not factored in this assessment. The proposed models namely DNN, RNN, CNN, and RNN+CNN give an accuracy of 55%, 61%, 55% and 98% respectively. Novelty and applications: The RNN+CNN model proposed to have good accuracy which is much better compared to existing models. The model in addition can be used in real-time CCTV surveillance which can predict the fire in real-time with a faster alert duration of less than 30 sec. Keywords: Fire detection, CNN, RNN, DNN, Artificial Intelligence (AI), Environment loss, Fire Losses

In credit scoring, data often has class-imbalanced problems. However, traditional cost-sensitive learning methods rarely consider the varying costs among samples. Moreover, previous studies have limitations, such as the lack of fit to real-world business needs and limited model interpretability. To address these issues, this paper proposes a novel example-dependent cost-sensitive learning based selective deep ensemble (ECS-SDE) model for customer credit scoring, which integrates example-dependent cost-sensitive learning with the interpretable TabNet (attentive interpretable tabular learning) and GMDH (group method of data handling) deep neural networks. Specifically, we use TabNet, which excels in handling tabular data, as the base classifier and optimize its performance on imbalanced data with an example-dependent cost loss function. Next, we design a GMDH based on an example-dependent cost-sensitive symmetric criterion to selectively deep integrate the base classifiers. This approach reduces the redundancy of base models in traditional ensemble strategies and enhances classification performance. Experimental results show that the ECS-SDE model outperforms six cost-sensitive models and five advanced deep ensemble models in overall performance for credit scoring. It shows significant advantages in the BS+, Save, and AUC metrics on four datasets. Furthermore, the ECS-SDE model provides strong interpretability, and detailed analysis reveals the key roles of various features in credit scoring.

<p>Cotton plant (Gossypium herbaceum), is one of the significant fiber crop grown worldwide. However, the crop is quite prone to leaves diseases, for which deep learning (DL) techniques can be utilized for early disease prediction and prevent stakeholders from losing the harvest. The objective of this paper is to develop a novel ensemble based deep convolutional neural network (DCNN) model developed on two base pretrained models named: VGG16 and InceptionV3 for early detection of cotton leaves diseases. The proposed ensemble model trained on cotton leaves dataset reports higher training and testing prediction accuracies as compared to the base pretrained models. Given that, deep learning architectures have hyper-parameters, this paper presents exhaustive experimental evaluations on ensemble model to tune hyper-parameters named learning rate, optimizer and no of epochs. The suggested hyper-parameter settings can be directly utilized while employing the ensemble model for cotton plant leaves disease detection and prediction. With suggested hyper-parameters settings of learning rate 0.0001, 20 epochs and stochastic gradient descent (SGD) optimizer, ensemble model reported training and testing accuracies of 98% and 95% respectively, which was higher than the training and testing accuracies of VGG16 and InceptionV3 pretrained DCNN models.</p>