Figure 7 - uploaded by Jinhai Xiang
Content may be subject to copyright.
Source publication
Although Generative Adversarial Networks (GAN) have shown remarkable performance in image generation, there exist some challenges in instability and convergence speed. During the training, the results of some models display the imbalances of quality within a generated image, in which some defective parts appear compared with other regions. Differen...
Similar publications
Room reverberation alters the spatial impression and timbre of a sound by modulating its spectral and temporal characteristics. Thus, we argue that, on a perceptual level, reverberation basically breaks down into interaural differences and spectro-temporal cues and that the separation of a perceived timbre into a sound source and a surrounding room...
Citations
... BigGAN was implemented on top of SAGAN architecture by employing certain techniques (such as truncation trick, orthogonal regularization), which substantially improves the performance of class-conditional GAN [28]. Liu et al. proposed an adaptive global and local bilevel optimization model (GL-GAN), that embedded a local bilevel optimization technique to improve the poor quality portion on an image along with traditional global optimization technique to optimize the whole image [29]. ...
Plant phenology and phenotype prediction using remote sensing data is increasingly gaining the attention of the plant science community to improve agricultural productivity. In this work, we generate synthetic forestry images that satisfy certain phenotypic attributes, viz. canopy greenness. The greenness index of plants describes a particular vegetation type in a mixed forest. Our objective is to develop a Generative Adversarial Network (GAN) to synthesize forestry images conditioned on this continuous attribute, i.e., greenness of vegetation, over a specific region of interest. The training data is based on the automated digital camera imagery provided by the National Ecological Observatory Network (NEON) and processed by the PhenoCam Network. The synthetic images generated by our method are also used to predict another phenotypic attribute, viz., redness of plants. The Structural SIMilarity (SSIM) index is utilized to assess the quality of the synthetic images. The greenness and redness indices of the generated synthetic images are compared against that of the original images using Root Mean Squared Error (RMSE) in order to evaluate their accuracy and integrity. Moreover, the generalizability and scalability of our proposed GAN model is determined by effectively transforming it to generate synthetic images for other forest sites and vegetation types.
... These challenges range from general problems with GANs to specific problems associated with histopathology samples. The general challenges coves issues relating to the instability of the model and reduced convergence speed [14]. Those widely identified problems of GAN for medical images are the increasing need to attenuate GANs to address noise removal from images, accurately segmenting samples, styling and reconstructing the new image from an actual image, and data simulation [15]. ...
... Similarly, the work in [27] promoted the use of metaheuristics as an optimization method for GANs, and the study in [28] also addressed the challenge of stabilizing GAN through the use of a genetic algorithm (GA). The authors in [14] solved the problem of instability and lack of early convergence by using a model which allows for global and local optimization methods. The global residual (GR) method [29] was proposed for stabilizing GANs. ...
... The issue of GAN optimization has been considered in [14] to address the problem of instability and lack of early convergence. Using a model which allows for global and local optimization methods, a DCGAN was adopted to build the new GAN model, namely (GL-GAN). ...
Generative adversarial networks (GAN) represent two deep learning (DL) models positioned in an adversarial manner to generate and evaluate images. This area of research promises to address several issues associated with medical image analysis using deep learning architectures and has been applied to medical image synthesis. The histopathology image samples are the gold standard for detecting and staging cancer since they contain rich latent information. However, this medical image modality is highly problematic: imbalanced class distribution in datasets, the rareness of publicly accessible pathologic findings, the burdensome task of image annotation, the increasing need to anonymize pathology samples, segmentation of regions of interest, and the demand for high-quality with super-resolution histopathology images. In this paper, we present a highly optimized, locally attenuated and two-level optimization strategy to improve the performance of GAN. First, a novel feature space-to-latent space mapping mechanism is designed to enrich the latent space of input to the generator. We applied a DL model to extract discriminant features and used dimensionality reduction to match the number of features to latent space. Secondly, a new metaheuristic algorithm, the Ebola optimization search algorithm (EOSA), optimizes the EOSA-GAN architecture and is experimentally applied to benchmark datasets. Results obtained showed that the quality of generated samples achieved an impressive outcome when evaluated using the Feature Similarity Indexing Method (FSIM), Peak Signal to Noise Ratio (PSNR), Structured Similarity Indexing Method (SSIM) and others. The finding from this study demonstrates the impact of optimization algorithms in stabilizing and speeding up GANs to convergence.
... 1. Generation of high-quality images Recent studies on GAN have enhanced both the usability and quality of image production abilities, such as the LAPGAN model [354] discussed Before. Several publications have addressed the issue of lack of training data using GANs [350,[362][363][364]. ...
Data scarcity is a major challenge when training deep learning (DL) models. DL demands a large amount of data to achieve exceptional performance. Unfortunately, many applications have small or inadequate data to train DL frameworks. Usually, manual labeling is needed to provide labeled data, which typically involves human annotators with a vast background of knowledge. This annotation process is costly, time-consuming, and error-prone. Usually, every DL framework is fed by a significant amount of labeled data to automatically learn representations. Ultimately, a larger amount of data would generate a better DL model and its performance is also application dependent. This issue is the main barrier for many applications dismissing the use of DL. Having sufficient data is the first step toward any successful and trustworthy DL application. This paper presents a holistic survey on state-of-the-art techniques to deal with training DL models to overcome three challenges including small, imbalanced datasets, and lack of generalization. This survey starts by listing the learning techniques. Next, the types of DL architectures are introduced. After that, state-of-the-art solutions to address the issue of lack of training data are listed, such as Transfer Learning (TL), Self-Supervised Learning (SSL), Generative Adversarial Networks (GANs), Model Architecture (MA), Physics-Informed Neural Network (PINN), and Deep Synthetic Minority Oversampling Technique (DeepSMOTE). Then, these solutions were followed by some related tips about data acquisition needed prior to training purposes, as well as recommendations for ensuring the trustworthiness of the training dataset. The survey ends with a list of applications that suffer from data scarcity, several alternatives are proposed in order to generate more data in each application including Electromagnetic Imaging (EMI), Civil Structural Health Monitoring, Medical imaging, Meteorology, Wireless Communications, Fluid Mechanics, Microelectromechanical system, and Cybersecurity. To the best of the authors’ knowledge, this is the first review that offers a comprehensive overview on strategies to tackle data scarcity in DL.
... In data generation algorithms, the synthetic minority over-sampling technique(SMOTE) (Chawla et al, 2002) implements interpolate operation to generate synthetic samples and is widely extended to other over-sampling algorithms (Camacho et al, 2022;Chen et al, 2022). Deep generation models like generative adversarial networks (GAN) (Goodfellow et al, 2020;Liu et al, 2022) and variational autoencoder (Kingma and Welling, 2014;Huang and Wang, 2022) are proposed to generate synthetic samples. GAN iteratively learns the hidden layer distribution of data through the generator and discriminator. ...
In graph neural network applications,GraphSAGE applies inductive learning and has been widely applied in important research topics such as node classification.The subgraph of nodes directly affects the classification performance for GraphSAGE due to it applies aggregation function to obtain embedding from the neighbors' feature.In many practical applications, the uneven class distribution of nodes makes it difficult for graph neural network to fully learn the topology and attribute of the minority, which limits the classification performance.Aiming at the problem of imbalanced node classification in GraphSAGE,we propose a new graph over-sampling algorithm called Subgraph Generation by Conditional Generative Adversarial Network (SG-CGAN).SG-CGAN learns the hidden layer expression of different nodes through GraphSAGEand trains conditional generative adversarial network(CGAN) through the nodes' hidden vector and related subgraph.Meanwhile, the hidden synthetic data is generated as input of CGAN to generate subgraphs of the minority,and retrain the GraphSAGE by adding the synthetic subgraphs.Experiments based on five graph datasets show that SG-CGAN can help GraphSAGE effectively improve ACC, macro-F1 and micro-F1,verifying the effectiveness of SG-CGAN generated data.
... Despite this great success, the standard GAN faces its challenges [14,17,37], and thus reaps various variants by taking different strategies [31,38]. Inherited from the classical repair models [16], we further consider that the training of GAIN lacks an adaptive process that dynamically adjusts its number of iterations. ...
Making reliable recovery of missing traffic data facilitates diverse applications of data-driven intelligent transportation system. But faced with correlation and heterogeneity along spatial–temporal dimensions, most existing works lack sufficient capability to capture these complex properties, resulting in suboptimal imputation performance. By addressing this challenge, we propose a hybrid framework TFs-DGAN consisting of dynamic adaptive generative adversarial networks (DA-GAN) with multi-view temporal factorizations (TFs), which can efficiently repair missing data by modeling those spatial–temporal correlations. Of these, DA-GAN model can generate traffic data from noise distribution and keep iterating dynamically to extract the global consistency. To further exploit the local consistency, TFs model drives the continual reduction in local elements in residuals by a novel truncation mechanism. Unlike the single model computation, TFs-DGAN integrates all stage-optimized residuals by local feedback and finally outputs the best repair results. In fact, our intention for this strategy is that DA-GAN module produces data but inaccurately, while TFs module refines its imperfections by modeling multi-view temporal properties. From the numerical analysis, the empirical evaluation relying on two publicly available traffic datasets suggests that our TFs-DGAN significantly outperforms the state-of-the-art baseline models in terms of accuracy, stability and computational efficiency.
... An adaptive global and local bilevel optimization model (GL-GAN) proposed in [42] optimized the image from the local and global aspects. The local bilevel optimization model was proposed based on the discriminator's output feature matrix, in which each element evaluates image receptive field quality and determines the area with low quality. ...
... Instead, the task drives D to focus only on small regions because the model can find class cues from small regions of the images. Focusing on limited regions or patterns is a typical overfitting behavior widely occurring for D in vanilla GANs.Furthermore, we compared LiWGAN with the state-of-art methods, including GL-GAN[42], DAG + Dist-GAN[61], ...
Generative adversarial networks (GANs) gained tremendous growth due to its potential and efficacy in producing realistic samples. This study proposes a light-weight GAN (LiWGAN) to learn the non-image synthesis with minimum computational time for less power computing. Hence, the LiWGAN method enhanced a new skip-layer channel-wise excitation module (SLE) and a self-supervised discriminator design for the non-synthesis performance using the facemask dataset. The facemask is one of the preventative strategies pioneered by the current COVID-19 pandemic. LiWGAN manipulates a non-image synthesis of facemask that could be beneficial for some researchers to identify an individual using lower power devices, occlusion challenges for face recognition, and alleviate the accuracy challenges due to limited datasets. The performance compared the processing time for a facemask dataset in terms of batch sizes and image resolutions. The Fréchet inception distance (FID) was also measured on the facemask images to evaluate the quality of the augmented image using LiWGAN. The findings for 3000 generated images showed a nearly similar FID score at 220.43 with significantly less processing time per iteration at 1.03s than StyleGAN at 219.97 FID score. One experiment was conducted using the CelebA dataset to compare with GL-GAN and DRAGAN, proving LiWGAN is appropriate for other datasets. The outcomes found LiWGAN performed better than GL-GAN and DRAGAN at 91.31 FID score with 3.50s processing time per iteration. Therefore, LiWGAN could aim to enhance the FID score to be near zero in the future with less processing time by using different datasets.
In machine learning, a generative model is responsible for generating new samples of data in terms of a probabilistic model. Generative adversarial network (GAN) has been widely used to generate realistic samples in different domains and outperforms its peers in the generative models family. However, producing a robust GAN model is not a trivial task because many challenges face the GAN during the training process and impact its performance, affecting the quality and diversity of the generated samples. In this article, we conduct a comprehensive review of GANs to present the fundamentals of GAN, including its components, types, and objective functions. Also, we present an overview of the evaluation matrices used to evaluate GAN models. Moreover, we list the applications of GANs and research work in various domains. Finally, we present the challenges that face GANs and highlight two significant issues, representing mode collapse and training instability, in addition to those research efforts that tackle these challenges.
This article is categorized under: Statistical Learning and Exploratory Methods of the Data Sciences > Deep Learning
Statistical Learning and Exploratory Methods of the Data Sciences > Neural Networks
