Generative Adversarial Network Model Research Articles

DAE-GAN: Underwater Image Super-Resolution Based on Symmetric Degradation Attention Enhanced Generative Adversarial Network

Underwater images often exhibit detail blurring and color distortion due to light scattering, impurities, and other influences, obscuring essential textures and details. This presents a challenge for existing super-resolution techniques in identifying and extracting effective features, making high-quality reconstruction difficult. This research aims to innovate underwater image super-resolution technology to tackle this challenge. Initially, an underwater image degradation model was created by integrating random subsampling, Gaussian blur, mixed noise, and suspended particle simulation to generate a highly realistic synthetic dataset, thereby training the network to adapt to various degradation factors. Subsequently, to enhance the network’s capability to extract key features, improvements were made based on the symmetrically structured blind super-resolution generative adversarial network (BSRGAN) model architecture. An attention mechanism based on energy functions was introduced within the generator to assess the importance of each pixel, and a weighted fusion strategy of adversarial loss, reconstruction loss, and perceptual loss was utilized to improve the quality of image reconstruction. Experimental results demonstrated that the proposed method achieved significant improvements in peak signal-to-noise ratio (PSNR) and underwater image quality measure (UIQM) by 0.85 dB and 0.19, respectively, significantly enhancing the visual perception quality and indicating its feasibility in super-resolution applications.

Deep learning-based whole-body PSMA PET/CT attenuation correction utilizing Pix-2-Pix GAN.

Sequential PET/CT studies oncology patients can undergo during their treatment follow-up course is limited by radiation dosage. We propose an artificial intelligence (AI) tool to produce attenuation-corrected PET (AC-PET) images from non-attenuation-corrected PET (NAC-PET) images to reduce need for low-dose CT scans. A deep learning algorithm based on 2D Pix-2-Pix generative adversarial network (GAN) architecture was developed from paired AC-PET and NAC-PET images. 18F-DCFPyL PSMA PET-CT studies from 302 prostate cancer patients, split into training, validation, and testing cohorts (n = 183, 60, 59, respectively). Models were trained with two normalization strategies: Standard Uptake Value (SUV)-based and SUV-Nyul-based. Scan-level performance was evaluated by normalized mean square error (NMSE), mean absolute error (MAE), structural similarity index (SSIM), and peak signal-to-noise ratio (PSNR). Lesion-level analysis was performed in regions-of-interest prospectively from nuclear medicine physicians. SUV metrics were evaluated using intraclass correlation coefficient (ICC), repeatability coefficient (RC), and linear mixed-effects modeling. Median NMSE, MAE, SSIM, and PSNR were 13.26%, 3.59%, 0.891, and 26.82, respectively, in the independent test cohort. ICC for SUVmax and SUVmean were 0.88 and 0.89, which indicated a high correlation between original and AI-generated quantitative imaging markers. Lesion location, density (Hounsfield units), and lesion uptake were all shown to impact relative error in generated SUV metrics (all p < 0.05). The Pix-2-Pix GAN model for generating AC-PET demonstrates SUV metrics that highly correlate with original images. AI-generated PET images show clinical potential for reducing the need for CT scans for attenuation correction while preserving quantitative markers and image quality.

A comprehensive end-to-end computer vision framework for restoration and recognition of low-quality engineering drawings

The digitization of engineering drawings is crucial for efficient reuse, distribution, and archiving. Existing computer vision approaches for digitizing engineering drawings typically assume the input drawings have high quality. However, in reality, engineering drawings are often blurred and distorted due to improper scanning, storage, and transmission, which may jeopardize the effectiveness of existing approaches. This paper focuses on restoring and recognizing low-quality engineering drawings, where an end-to-end framework is proposed to improve the quality of the drawings and identify the graphical symbols on them. The framework uses K-means clustering to classify different engineering drawing patches into simple and complex texture patches based on their gray level co-occurrence matrix statistics. Computer vision operations and a modified Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) model are then used to improve the quality of the two types of patches, respectively. A modified Faster Region-based Convolutional Neural Network (Faster R-CNN) model is used to recognize the quality-enhanced graphical symbols. Additionally, a multi-stage task-driven collaborative learning strategy is proposed to train the modified ESRGAN and Faster R-CNN models to improve the resolution of engineering drawings in the direction that facilitates graphical symbol recognition, rather than human visual perception. A synthetic data generation method is also proposed to construct quality-degraded samples for training the framework. Experiments on real-world electrical diagrams show that the proposed framework achieves an accuracy of 98.98% and a recall of 99.33%, demonstrating its superiority over previous approaches. Moreover, the framework is integrated into a widely-used power system software application to showcase its practicality. The reference codes and data can be found at https://github.com/Lattle-y/AI-recognition-for-lq-ed.git Future work will focus on improving the generalizability of the proposed framework to different quality degradation scenarios and extrapolating the application to different engineering domains.

In-Domain GAN Inversion for Faithful Reconstruction and Editability.

Generative Adversarial Networks (GANs) have significantly advanced image synthesis through mapping randomly sampled latent codes to high-fidelity synthesized images. However, applying well-trained GANs to real image editing remains challenging. A common solution is to find an approximate latent code that can adequately recover the input image to edit, which is also known as GAN inversion. To invert a GAN model, prior works typically focus on reconstructing the target image at the pixel level, yet few studies are conducted on whether the inverted result can well support manipulation at the semantic level. This work fills in this gap by proposing in-domain GAN inversion, which consists of a domain-guided encoder and a domain-regularized optimizer, to regularize the inverted code in the native latent space of the pre-trained GAN model. In this way, we manage to sufficiently reuse the knowledge learned by GANs for image reconstruction, facilitating a wide range of editing applications without any retraining. We further make comprehensive analyses on the effects of the encoder structure, the starting inversion point, as well as the inversion parameter space, and observe the trade-off between the reconstruction quality and the editing property. Such a trade-off sheds light on how a GAN model represents an image with various semantics encoded in the learned latent distribution.

Decision-making dynamics in urban energy transition: A statistical analysis of policy frameworks, stakeholder collaboration and technological innovations

This research explores the intricate dynamics of decision-making in the context of urban energy transition, focusing on the interplay between policy frameworks, stakeholder collaboration, and technological innovations. The study makes three key contributions to the field. Firstly, it proposes an innovative approach to enhance market price prediction accuracy by integrating a Generative Adversarial Network (GAN) model. Leveraging the Adaptive Bat Algorithm (ABA) for GAN parameter optimization, the model exhibits improved precision in forecasting electricity prices, thereby providing stakeholders with more reliable insights. Secondly, the research introduces optimized model parameters for efficient energy decision-making, employing the ABA to dynamically adapt GAN parameters. This adaptive process enhances the overall performance of the model, facilitating better-informed decisions for stakeholders in the energy sector. Lastly, the integration of the ABA for GAN parameter optimization contributes to informed decision-making by generating more accurate market price values. This precision empowers policymakers, utilities, and investors with reliable insights, facilitating the development and implementation of strategies aligned with the goals of sustainable and efficient urban energy transition. The research underscores the significance of accurate market price predictions in shaping a resilient and sustainable energy landscape.

Transformer fault diagnosis based on adversarial generative networks and deep stacked autoencoder

Establishing a deep learning model for transformer fault diagnosis using transformer oil chromatogram data requires a large number of fault samples. The lack and imbalance of oil chromatogram data can lead to overfitting, lack of representativeness of the model, and unsatisfactory prediction results on test set data, making it difficult to accurately diagnose transformer faults. A conditional Wasserstein generative adversarial network with gradient penalty optimization (CWGAN-GP) is adopted in this paper, which based on gradient penalty optimization and expand the oil chromatography fault samples of 500 sets of transformer oil chromatography data with 5 types of faults. The proposed method is used to classify transformer faults using a deep autoencoder, and the sample quality of the neural network model proposed in this paper is compared with several other variants of generative adversarial neural network models. The research results show that after using the method proposed in this paper for sample expansion, the overall accuracy of fault diagnosis can reach 93.2 %, which is 4.98 % higher than the original imbalanced samples. Compared with other sample expansion methods, the accuracy of fault diagnosis of the algorithm in this paper is improved by 1.70 %–3.05 %.

Vein Pattern Visualisation for Biometric Identification with cGAN on a New Zealand Dataset

Forensic identification using vein patterns in standard colour images presents significant challenges due to their low visibility. Recent efforts have employed various computational techniques, including artificial neural networks and optical vein disclosure, to enhance vein pattern detection. However, these methods still face limitations in reliability when compared to Near-Infrared (NIR) reference images. One of the biggest challenges of the studies is the limited number of available datasets that have synchronised colour and NIR images from body limbs. This paper introduces a new dataset comprising 602 pairs of synchronised NIR and RGB forearm images from a diverse population, ethically approved and collected in Auckland, New Zealand. Using this dataset, we also propose a conditional Generative Adversarial Networks (cGANs) model to translate RGB images into their NIR equivalents. Our evaluations focus on matching accuracy, vein length measurements, and contrast quality, demonstrating that the translated vein patterns closely resemble their NIR counterparts. This advancement offers promising implications for forensic identification techniques.

Text Feature Adversarial Learning for Text Generation With Knowledge Transfer From GPT2.

Text generation is a key component of many natural language tasks. Motivated by the success of generative adversarial networks (GANs) for image generation, many text-specific GANs have been proposed. However, due to the discrete nature of text, these text GANs often use reinforcement learning (RL) or continuous relaxations to calculate gradients during learning, leading to high-variance or biased estimation. Furthermore, the existing text GANs often suffer from mode collapse (i.e., they have limited generative diversity). To tackle these problems, we propose a new text GAN model named text feature GAN (TFGAN), where adversarial learning is performed in a continuous text feature space. In the adversarial game, GPT2 provides the "true" features, while the generator of TFGAN learns from them. TFGAN is trained by maximum likelihood estimation on text space and adversarial learning on text feature space, effectively combining them into a single objective, while alleviating mode collapse. TFGAN achieves appealing performance in text generation tasks, and it can also be used as a flexible framework for learning text representations.

Implementing generative adversarial networks for increasing performance of transmission fault classification

&lt;p&gt;An electrical power system is a network that facilitates the sourcing, transfer, and distribution of electrical energy. In the traditional power system, there are eleven types of faults that can occur in the system. This paper focuses on the classification of these faults over a stretch of 100 kilometres. The dataset used is synthetic and generated from a simulated model using MATLAB/Simulink software. Data augmentation is carried out during training to improve the accuracy of the classification. An indirect training approach through generative adversarial network (GAN) is used to classify these overhead transmission line faults. The random forest (RF) classification is used as the base learning model on the original dataset and it achieves accuracy of 84%. However, the base learner RF when used on GAN model generated augmented faulty data, it performs exceptionally well achieving 99% accuracy. One of the recent state-of-art methods is compared with this approach.&lt;/p&gt;

An optimal framework for natural English translate processing: On the application of artificial neural network and Adam optimization algorithm

Traditional rule-based and statistical methods have limitations when dealing with complex language structures and semantics. In neural network machine translation algorithms, the objective function is usually to improve the accuracy of n-ary words. However, this does not guarantee a more natural and accurate translation. To overcome these challenges, this paper proposes an optimization algorithm for English natural translation processing based on neural networks, which combines Generative Adversarial Network (GAN) and Transformer models. In GAN, the generative model uses the Transformer model to generate false samples, while the discriminative model uses a binary classifier based on convolutional neural networks and attention mechanisms to distinguish between true and false samples. During the training process, reinforcement learning algorithms are added to evaluate and adjust the generated sentences, and the parameters of the generated model are updated. The classification results of the discriminative model are used together with the Bilingual Evaluation Basis Value (BLEU) objective function to evaluate false samples, and the results are fed back to the generating model to guide parameter updates and optimization. Extensive experiments were conducted on a standard English-Chinese machine translation dataset to evaluate our method. Compared with the benchmark model that only uses supervised learning methods, our neural network-based optimization algorithm for English natural translation processing has achieved significant improvements in translation quality. According to statistical comparison, compared with the Transformer model (BLUE = 33.63 and AP = 90%) and the deep learning model based on long-term and short-term memory (BLUE = 30.26 and AP = 83%), the GAN and Transformer models proposed as the best framework exhibit better performance in bilingual evaluation deficiency (BLEU) (34.35) and accuracy (AP = 95%).

Prediction of Myocardial Infarction Using a Combined Generative Adversarial Network Model and Feature-Enhanced Loss Function

Accurate risk prediction for myocardial infarction (MI) is crucial for preventive strategies, given its significant impact on global mortality and morbidity. Here, we propose a novel deep-learning approach to enhance the prediction of incident MI cases by incorporating metabolomics alongside clinical risk factors. We utilized data from the KORA cohort, including the baseline S4 and follow-up F4 studies, consisting of 1454 participants without prior history of MI. The dataset comprised 19 clinical variables and 363 metabolites. Due to the imbalanced nature of the dataset (78 observed MI cases and 1376 non-MI individuals), we employed a generative adversarial network (GAN) model to generate new incident cases, augmenting the dataset and improving feature representation. To predict MI, we further utilized multi-layer perceptron (MLP) models in conjunction with the synthetic minority oversampling technique (SMOTE) and edited nearest neighbor (ENN) methods to address overfitting and underfitting issues, particularly when dealing with imbalanced datasets. To enhance prediction accuracy, we propose a novel GAN for feature-enhanced (GFE) loss function. The GFE loss function resulted in an approximate 2% improvement in prediction accuracy, yielding a final accuracy of 70%. Furthermore, we evaluated the contribution of each clinical variable and metabolite to the predictive model and identified the 10 most significant variables, including glucose tolerance, sex, and physical activity. This is the first study to construct a deep-learning approach for producing 7-year MI predictions using the newly proposed loss function. Our findings demonstrate the promising potential of our technique in identifying novel biomarkers for MI prediction.

DRL-GAN: A Hybrid Approach for Binary and Multiclass Network Intrusion Detection.

Our increasingly connected world continues to face an ever-growing number of network-based attacks. An Intrusion Detection System (IDS) is an essential security technology used for detecting these attacks. Although numerous Machine Learning-based IDSs have been proposed for the detection of malicious network traffic, the majority have difficulty properly detecting and classifying the more uncommon attack types. In this paper, we implement a novel hybrid technique using synthetic data produced by a Generative Adversarial Network (GAN) to use as input for training a Deep Reinforcement Learning (DRL) model. Our GAN model is trained on the NSL-KDD dataset, a publicly available collection of labeled network traffic data specifically designed to support the evaluation and benchmarking of IDSs. Ultimately, our findings demonstrate that training the DRL model on synthetic datasets generated by specific GAN models can result in better performance in correctly classifying minority classes over training on the true imbalanced dataset.

Global interactive-mode imperfection generation for K6 single-layer latticed shell using generative adversarial networks

The worst imperfection shape of K6 single-layer latticed shells corresponding to the lowest nonlinear load-carrying capacity is difficult to be obtained in design owing to significant multi-modal interaction. The global interactive buckling of the K6 single-layer latticed shell is investigated numerically, and the intelligent model for generating its worst global interactive-mode imperfection is developed via generative adversarial networks (GAN). For the cases vulnerable to interactive buckling, the load-carrying capacity is found to be reduced significantly up to 35% with two-mode interaction considered, and unstable behaviour, such as the dominant mode changing due to the deformation evolution of different modal components, is observed at the post-buckling stage. Moreover, the GAN model for generating the worst imperfection shape and the artificial neural network for evaluating the corresponding ultimate load are developed based on the data obtained from numerical analysis. The time of training model to converge is less than 30 min, and the generating process using the trained model only takes a few seconds. It is demonstrated that for the K6 single-layer latticed shells, the developed models can give the worst imperfection and the corresponding ultimate load accurately and efficiently with interactive buckling considered well. This work can be used to develop the programming module for the intelligent design of latticed shells in future.

Analyzing microstructure relationships in porous copper using a multi-method machine learning-based approach

The prediction of material properties from a given microstructure and its reverse engineering displays an essential ingredient for accelerated material design. However, a comprehensive methodology to uncover the processing-structure-property relationship is still lacking. Herein, we develop a methodology capable of understanding this relationship for differently processed porous materials. We utilize a multi-method machine learning approach incorporating tomographic image data acquisition, segmentation, microstructure feature extraction, feature importance analysis and synthetic microstructure reconstruction. Enhanced segmentation with an accuracy of about 95% based on an efficient annotation technique provides the basis for accurate microstructure quantification, prediction and understanding of the correlation of the extracted microstructure features and electrical conductivity. We show that a diffusion probabilistic model superior to a generative adversarial network model, provides synthetic microstructure images including physical information in agreement with real data, an essential step to predicting properties of unseen conditions.

Image Enhancement Based on Dual-Branch Generative Adversarial Network Combining Spatial and Frequency Domain Information for Imbalanced Fault Diagnosis of Rolling Bearing

To address the problems of existing 2D image-based imbalanced fault diagnosis methods for rolling bearings, which generate images with inadequate texture details and color degradation, this paper proposes a novel image enhancement model based on a dual-branch generative adversarial network (GAN) combining spatial and frequency domain information for an imbalanced fault diagnosis of rolling bearing. Firstly, the original vibration signals are converted into 2D time–frequency (TF) images by a continuous wavelet transform, and a dual-branch GAN model with a symmetric structure is constructed. One branch utilizes an auxiliary classification GAN (ACGAN) to process the spatial information of the TF images, while the other employs a GAN with a frequency generator and a frequency discriminator to handle the frequency information of the input images after a fast Fourier transform. Then, a shuffle attention (SA) module based on an attention mechanism is integrated into the proposed model to improve the network’s expression ability and reduce the computational burden. Simultaneously, mean square error (MSE) is integrated into the loss functions of both generators to enhance the consistency of frequency information for the generated images. Additionally, a Wasserstein distance and gradient penalty are also incorporated into the losses of the two discriminators to prevent gradient vanishing and mode collapse. Under the supervision of the frequency WGAN-GP branch, an ACWGAN-GP can generate high-quality fault samples to balance the dataset. Finally, the balanced dataset is utilized to train the auxiliary classifier to achieve fault diagnosis. The effectiveness of the proposed method is validated by two rolling bearing datasets. When the imbalanced ratios of the four datasets are 0.5, 0.2, 0.1, and 0.05, respectively, their average classification accuracy reaches 99.35% on the CWRU bearing dataset. Meanwhile, the average classification accuracy reaches 96.62% on the MFS bearing dataset.

Oil spill detection and classification through deep learning and tailored data augmentation

Oil spill detection is an important task for protecting and minimizing the harmful effects of oil on the marine ecosystem. Currently, the application of images from unmanned aerial vehicles, along with deep learning, is widely employed. Although these methods have yielded good results, the issue of oil spill classification based on these methods has not received much attention at present. In this research, a deep learning model with a dual attention mechanism consisting of two modules was utilized. The first module focuses on capturing the spatial relationships between each pixel and the entire image, the second module identifies the characteristics between channels in the image, thereby enhancing the ability to detect and classify oil. Additionally, a data augmentation technique based on the Generative Adversarial Networks model was refined and employed to improve the model's accuracy. Experimental results, obtained through comparisons between dataset construction methods, the use of different encoders and decoders, and adjustments hyperparameters, reveal that the best model achieves a mean Intersection over Union by 72.49%. Data augmentation techniques also contribute to a 2.56% increase in mean Intersection over Union. The findings of this research provide a feasible solution not only for detecting but also for classifying oil spills, thereby assisting marine environmental managers in making timely decisions to respond to oil spill accidents.

Structural seismic response reconstruction method based on multidomain feature-guided generative adversarial neural networks

Structural seismic response reconstruction is important to assess the safety of structures. This study presents a novel multidomain feature-guided generative adversarial neural network model (MWGAN-TF) for reconstructing the seismic responses of structures, which takes into account the joint non-stationarity of the seismic response in the time-frequency statistical domain. It innovatively incorporates time, frequency, and statistical-domain feature constraints into the multiscale generative adversarial neural network, which guides the model to learn the multidomain feature information of the seismic response at different time scales. A statistical indicator (CNCSI) was proposed to evaluate the performance of the model in capturing nonstationary characteristics. The effectiveness of the MWGAN-TF was verified using response data from numerical models of a three-story moment-resisting frame and reinforced concrete frame structures, as well as the field measurement data of an actual building. Thereafter, the effects of different domain feature-guided models on the reconstruction response accuracy are discussed. The results show that embedding multidomain feature constraints can provide a more reliable response reconstruction by improving the ability of the model to capture nonstationary characteristics. Thus, the deep learning paradigm based on multidomain feature guidance outperforms the classical neural network guided only by time-domain features.

Multilingual Hate Speech Detection: A Semi-Supervised Generative Adversarial Approach.

Social media platforms have surpassed cultural and linguistic boundaries, thus enabling online communication worldwide. However, the expanded use of various languages has intensified the challenge of online detection of hate speech content. Despite the release of multiple Natural Language Processing (NLP) solutions implementing cutting-edge machine learning techniques, the scarcity of data, especially labeled data, remains a considerable obstacle, which further requires the use of semisupervised approaches along with Generative Artificial Intelligence (Generative AI) techniques. This paper introduces an innovative approach, a multilingual semisupervised model combining Generative Adversarial Networks (GANs) and Pretrained Language Models (PLMs), more precisely mBERT and XLM-RoBERTa. Our approach proves its effectiveness in the detection of hate speech and offensive language in Indo-European languages (in English, German, and Hindi) when employing only 20% annotated data from the HASOC2019 dataset, thereby presenting significantly high performances in each of multilingual, zero-shot crosslingual, and monolingual training scenarios. Our study provides a robust mBERT-based semisupervised GAN model (SS-GAN-mBERT) that outperformed the XLM-RoBERTa-based model (SS-GAN-XLM) and reached an average F1 score boost of 9.23% and an accuracy increase of 5.75% over the baseline semisupervised mBERT model.

Novel virtual sensors development based on machine learning combined with convolutional neural-network image processing-translation for feedback control systems of internal combustion engines

Physical sensors are commonly used to record performance data of internal combustion engines (ICEs) for online feedback control and calibration, but they are prone to diagnostic and increased development costs. Lookup tables are commonly used in conventional calibration and feedback control; however, the table parameters increase with the advancement of ICE technologies under transient operations. Consequently, the calibration and control systems are time-consuming. This work proposes novel virtual sensors to address these issues by predicting the combustion, performance, and emission of ICEs using neural networks and image processing/translation. The novel sensors are targeted for onboard feedback control systems under transient driving. Firstly, a virtual diesel engine (VDE) was developed and calibrated against experimental data taken from a production 2.2 L turbocharged diesel engine. The VDE was calibrated under WLTC, JC08, and NEDC transient operations and was used to generate teaching data. Next, the virtual sensors are developed using five machine learning (ML) regressors. The result shows that the coefficient of determination R2 from all ML regressors exceeded 0.94, and the XG-Boost outperforms other ML techniques with R2 > 0.977. XG-Boost parameter estimations were 8 times faster than that on a desktop simulation. Then, an image classification model using a deep convolutional neural network (D-CNN) is constructed, and the dependency of performance parameters and exhaust emissions with the rate of heat release (R.H.R) and in-cylinder pressure profile is confirmed. The performance parameters and emissions dependency was compared individually with R.H.R. and the in-cylinder pressure profile. As a result, a strong correlation between the performance and R.H.R. was observed. Finally, a generative adversarial network (GAN) model was constructed to translate the in-cylinder pressure profile to R.H.R. profile. A novel method to develop virtual sensors for advanced feedback control of any type of ICEs is proposed for the first time.

AP-GAN-DNN based creep fracture life prediction for 7050 aluminum alloy

7050 aluminum alloy has high strength and excellent fracture toughness. However, damage can be caused to the 7050 aluminum alloy by creep at elevated temperatures. The accuracy of conventional creep life prediction methods may fall short of meeting the desired standards. And the application of machine learning (ML) methods is limited due to the scarcity and dispersion of creep sample data. Generative Adversarial Networks (GAN) are often used for data enhancement, but suffer from “pattern collapse” under small sample conditions. To solve the above problems, this paper proposes a creep fracture life prediction method, which combines the affinity propagation (AP) clustering algorithm, GAN, and deep neural networks (DNN) for accurate creep fracture life prediction. The AP clustering algorithm for adaptive clustering was utilized to better reflect the creep sample distribution. Independent GAN models were trained for each cluster to better capture distribution characteristics of the data and generate synthetic data that was highly similar to the real data. The DNN models were trained using the synthetic data and then predicted using real creep fracture life data. The presented method was compared with traditional physical and machine learning methods, and the method combining K-Means, GAN, and DNN. The experimental results show that the method proposed in this paper has better prediction accuracy in a small sample creep fracture life dataset.