Search

In an era of unprecedented technological advancement, the power industry is undergoing a transformative evolution, particularly in intelligent power grid enterprises. The integration of cutting-edge information science and technology has ushered in a new era of automation, informatization, and intelligence within these enterprises. While this progression promises enhanced production, operation, and management capabilities, it also brings forth a daunting challenge: the effective governance of the burgeoning volumes of power data. This paper aims to conduct in-depth research on evaluating the effectiveness of data governance in power grid enterprises based on deep learning to integrate more closely with their business systems. The main objective is to provide effective and convenient intelligent services for decision-making within an innovative power grid enterprise management system and strengthen the data architecture of these enterprises in data management. First, the deep learning neural network's principle structure and training methods are introduced in detail and combined with the deep learning neural network. This is a different evaluation model for the data governance effectiveness of power grid enterprises based on penalty variable weight. The difference probability density of the power difference data series in the power grid is taken as the evaluation index. The evaluation model for the governance effectiveness of different data is modified. Build a different evaluation model of power grid enterprise data governance effectiveness based on punishment and weight change, comprehensively consider the extent to which the data volume of power grid abnormal data in power grid enterprises affects the evaluation of data governance effectiveness, and complete the assessment of power grid enterprise data governance effectiveness based on deep learning. Experimental results underscore the method's efficacy, demonstrating an exceptional accuracy rate of 94%. This empirical validation highlights the method's efficient evaluation process, offering invaluable technical support for enhancing power data management's consistency, precision, and reliability within power grid enterprises. Moreover, comparative analyses against other methodologies, including KNN, SVM, RF, DT, and RNN, reaffirm the superiority of the DNN model, solidifying its outstanding performance. [ABSTRACT FROM AUTHOR]

The change detection (CD) technology has greatly improved the ability to interpret land surface changes. Deep learning (DL) methods have been widely used in the field of CD due to its high detection accuracy and application range. DL-based CD methods usually cannot fuse the extracted feature information at full scale, leaving out effective information, and commonly use transfer learning methods, which rely on the original dataset and training weights. To address the above issues, we propose a deeply supervised (DS) change detection network (DASUNet) that fuses full-scale features, which adopts a Siamese architecture, fuses full-scale feature information, and realizes end-to-end training. In order to obtain higher feature information, the network uses atrous spatial pyramid pooling (ASPP) module in the coding stage. In addition, the DS module is used in the decoding stage to exploit feature information at each scale in the final prediction. The experimental comparison shows that the proposed network has the current state-of-the-art performance on the CDD and the WHU-CD, reaching 94.32% and 90.37% on F1, respectively. [ABSTRACT FROM AUTHOR]

Purpose: In this paper, for the purpose of accurate and efficient mass detection, we propose a new deep learning framework, including two major stages: Suspicious region localization (SRL) and Multicontext Multitask Learning (MCMTL). Methods: In the first stage, SRL focuses on finding suspicious regions [regions of interest (ROIs)] and extracting multisize patches of these suspicious regions. A set of bounding boxes with different size is used to extract multisize patches, which aim to capture diverse context information. In the second stage, MCMTL networks integrate features from multisize patches of suspicious regions for classification and segmentation simultaneously, where the purpose of this stage is to keep the true positive suspicious regions and to reduce the false positive suspicious regions. Results: According to the experimental results on two public datasets (i.e., CBIS‐DDSM and INBreast), our method achieves the overall performance of 0.812 TPR@2.53 FPI and 0.919 TPR@0.12 FPI on test sets, respectively. Conclusions: Our proposed method suggests comparable performance to the state‐of‐the‐art methods. [ABSTRACT FROM AUTHOR]

Some recent studies use a convolutional neural network (CNN) or long short-term memory (LSTM) to extract gait features, but the methods based on the CNN and LSTM have a high loss rate of time-series and spatial information, respectively. Since gait has obvious time-series characteristics, while CNN only collects waveform characteristics, and only uses CNN for gait recognition, this leads to a certain lack of time-series characteristics. LSTM can collect time-series characteristics, but LSTM results in performance degradation when processing long sequences. However, using CNN can compress the length of feature vectors. In this paper, a sequential convolution LSTM network for gait recognition using multimodal wearable inertial sensors is proposed, which is called SConvLSTM. Based on 1D-CNN and a bidirectional LSTM network, the method can automatically extract features from the raw acceleration and gyroscope signals without a manual feature design. 1D-CNN is first used to extract the high-dimensional features of the inertial sensor signals. While retaining the time-series features of the data, the dimension of the features is expanded, and the length of the feature vectors is compressed. Then, the bidirectional LSTM network is used to extract the time-series features of the data. The proposed method uses fixed-length data frames as the input and does not require gait cycle detection, which avoids the impact of cycle detection errors on the recognition accuracy. We performed experiments on three public benchmark datasets: UCI-HAR, HuGaDB, and WISDM. The results show that SConvLSTM performs better than most of those reporting the best performance methods, at present, on the three datasets. [ABSTRACT FROM AUTHOR]

As a common technology in social network, clustering has attracted lots of research interest due to its high performance, and many clustering methods have been presented. The most of existing clustering methods are based on unsupervised learning. In fact, we usually can obtain some/few labeled samples in real applications. Recently, several semi-supervised clustering methods have been proposed, while there is still much space for improvement. In this paper, we aim to tackle two research questions in the process of semi-supervised clustering: (i) How to learn more discriminative feature representations to boost the process of the clustering; (ii) How to effectively make use of both the labeled and unlabeled data to enhance the performance of clustering. To address these two issues, we propose a novel semi-supervised clustering approach based on deep metric learning (SCDML) which leverages deep metric learning and semi-supervised learning effectively in a novel way. To make the extracted features of the contribution of data more representative and the label propagation network more suitable for real applications, we further improve our approach by adopting triplet loss in deep metric learning network and combining bedding with label propagation strategy to dynamically update the unlabeled to labeled data, which is named as semi-supervised clustering with deep metric learning and graph embedding (SCDMLGE). SCDMLGE enhances the robustness of metric learning network and promotes the accuracy of clustering. Substantial experimental results on Mnist, CIFAR-10, YaleB, and 20-Newsgroups benchmarks demonstrate the high effectiveness of our proposed approaches. [ABSTRACT FROM AUTHOR]

In the blast furnace (BF) ironmaking process, the gas utilization rate (GUR) is a crucial indicator for reflecting the energy consumption and operating status of BF. However, due to the complex and harsh environment in the BF top, accurately obtaining GUR online is not an effortless task. Although many studies have been carried out to predict GUR through data‐driven methods, some challenges still exist: 1) limited feature extraction capability for complex data patterns; 2) prediction accuracy is sensitive to the fluctuation of BF working conditions. Therefore, a novel deep learning method is proposed based on dynamic weighted stacked output‐relevant autoencoder (DW‐SOAE) for GUR online prediction. First, the input layer variables for each AE are weighted according to their importance, which will reduce output‐unrelated features. Then, the output variable is also reconstructed at the output layer of each AE, which ensures extracted features can largely predict GUR. Next, considering that the fluctuation of BF working conditions may affect prediction accuracy, the density peak clustering algorithm is used to cluster the process variables, and several DW‐SOAE‐based submodels are built for GUR prediction. Finally, the effectiveness and superiority of the proposed GUR prediction method are verified in industrial experiments. [ABSTRACT FROM AUTHOR]

Hashing has been widely used for large-scale remote sensing image retrieval due to its outstanding advantages in storage and search speed. Recently, deep hashing methods, which produce discriminative hash codes by building end-to-end deep convolutional networks, have shown promising results. However, training these networks requires numerous labeled images, which are scarce and expensive in remote sensing datasets. In order to solve this problem, we propose a deep unsupervised hashing method, namely deep contrastive self-supervised hashing (DCSH), which uses only unlabeled images to learn accurate hash codes. It eliminates the need for label annotation by maximizing the consistency of different views generated from the same image. More specifically, we assume that the hash codes generated from different views of the same image are similar, and those generated from different images are dissimilar. On the basis of the hypothesis, we can develop a novel loss function containing the temperature-scaled cross-entropy loss and the quantization loss to train the developed deep network end-to-end, resulting in hash codes with semantic similarity preserved. Our proposed network contains four parts. First, each image is transformed into two different views using data augmentation. After that, they are fed into an encoder with the same shared parameters to obtain deep discriminate features. Following this, a hash layer converts the high-dimensional image representations into compact binary codes. Lastly, a novel hash function is introduced to train the proposed network end-to-end and thus guide generated hash codes with semantic similarity. Extensive experiments on two popular benchmark datasets of the UC Merced Land Use Database and the Aerial Image Dataset have demonstrated that our DCSH has significant superiority in remote sensing image retrieval compared with state-of-the-art unsupervised hashing methods. [ABSTRACT FROM AUTHOR]

Many medical image datasets have been collected without proper annotations for deep learning training. In this paper, we propose a novel unsupervised domain adaptation framework with adversarial learning to minimize the annotation efforts. Our framework employs a task specific network, i.e. , fully convolutional network (FCN), for spatial density prediction. Moreover, we employ a domain discriminator, in which adversarial learning is adopted to align the less-annotated target domain features with the well-annotated source domain features in the feature space. We further propose a novel training strategy for the adversarial learning by coupling data from source and target domains and alternating the subnet updates. We employ the public CBIS-DDSM dataset as the source domain, and perform two sets of experiments on two target domains (i.e. , the public INbreast dataset and a self-collected dataset), respectively. Experimental results suggest consistent and comparable performance improvement over the state-of-the-art methods. Our proposed training strategy is also proved to converge much faster. [ABSTRACT FROM AUTHOR]

Abstract Pornographic image recognition is a challenging task due to the subjective definition and complex contextual information. In this paper, we propose an ensemble framework using uncertain inference (EFUI) for pornographic image recognition. The EFUI employs bayesian network (BN) as uncertain inference engine, while prior global confidence and uncertain evidence of local semantic components are acquired by deep learning networks. Specifically, we construct the graphical model of BN based on the internal contextual relationship of local semantic components, which conforms to common sense. The prior global confidence of pornography for candidate image is extracted using GoogleNet/ResNet-50. We extract the uncertain evidence of local semantic components, represented by the probability of visual object presence, using Single Shot MultiBox Detector (SSD). Finally, a novel uncertain belief propagation algorithm is introduced to propagate the belief of extracted uncertain evidence until convergence, then identifying the pornographic possibility of the image according to the final confidence. To evaluate EFUI, we employ the public NPDI Pornography database and collect a practical dataset contains 10 million images. Experimental results on all datasets demonstrate that our EFUI achieves the state-of-the-art performance on all evaluation metrics, and outperforms the best counterparts by up 6.95% and 0.61% for accuracy on the two datasets respectively. It also presents a comparable performance over the state-of-the-art approaches in literature. [ABSTRACT FROM AUTHOR]

• A novel debiased contrastive hashing method for accurate image retrieval. • Introducing neighborhood discovery to alleviate the bias problem caused by false negatives. • We bridge contrastive hashing and the EM algorithm to provide some insights. • Our proposed method outperforms state-of-the-art unsupervised hashing methods. Hashing has achieved great success in multimedia retrieval due to its high computing efficiency and low storage cost. Recently, contrastive-learning-based hashing methods have achieved decent retrieval performance in label-free scenarios by learning distortion-invariant representations with Siamese networks. Their learning principle, i.e., instance discrimination , maximizes the correlation between self-augmented views and treats all others as negative samples. However, it may learn with false negative samples that are naturally similar, resulting in biased hash learning. To bridge this flaw, we reveal the between-instance similarity of naturally similar samples by exploring the latent structure of the training data. As a result, we propose the D eep D ebiased C ontrastive H ashing (DDCH) algorithm, using the neighborhood discovery module to explore the intrinsic similarity relationship that can help contrastive hashing reduce false negatives for superior discriminatory ability. Furthermore, we elucidate the rationale for incorporating the module into the contrastive hashing framework and explain our hashing process from an Expectation-Maximization (EM) perspective. Extensive experimental results on three benchmark image datasets demonstrate that DDCH significantly outperforms the state-of-the-art unsupervised hashing methods for image retrieval. [ABSTRACT FROM AUTHOR]

Infrared imaging-based machine vision (IRMV) is the technology used to automatically inspect, detect, and analyse infrared images (or videos) obtained by recording the intensity of infrared light emitted or reflected by observed objects. Depending on whether controllable excitation is used during the imaging of infrared rays, thermal IRMV can be categorised into passive thermography and active thermography. Passive thermography is an important supplement to conventional machine vision based on visible light and is a valid imaging tool for self-heating objects such as the human body and electrical power devices. Active thermography is a non-destructive testing method for the quality evaluation and safety assurance of non-self-heating objects. In active thermography, the trend is to inspect rapidly, reliably, and intelligently by introducing multiple-mode excitation sources and artificial intelligence. The rapid development of deep learning makes IRMV more and more intelligent and highly automated, thus considerably increasing its range of applications. This paper reviews the principle, cameras, and thermal data of IRMV and discusses the applications of deep learning applied to IRMV. Case studies of IRMV and deep learning on various platforms such as unmanned vehicles, mobile phones and embedded systems are also reported. [ABSTRACT FROM AUTHOR]