LSUN-Stanford Car Dataset: Enhancing Large-Scale Car Image Datasets Using Deep Learning for Usage in GAN Training

Generative adversarial networks (GANs), which can generate and transform data, have been attracting attention. However, the model must be lightweight and fast when applied in the field. As for the ternarization of GAN, (TernaryGAN,) which restricts the value of the weights to {−1, 0, +1} during forwarding propagation of the generator, has already been proposed. In this paper, we investigated by experiment how ternary generator and/or discriminator affects the training of GANs. To make not only generator ternary, but also discriminator, we propose the DGR (Decomposition with Gradient Retained) method, which can change discriminator's input images to binary. We trained GANs for the cases where the generator and the discriminator are ternarized, and for the case where only one of them is ternarized, and measured the degree of image degradation using the FID (Fréchet inception distance) score. Only the ternarized generator is showed the lowest accuracy degradation, implying that GANs contain some parts that are not suitable for ternarization. We found the useful insight that when reducing the weight of GAN, the generator can be compressed relatively more, while the discriminator should not be so much.

BackgroundMalignant melanoma (MM) is the most aggressive skin cancer, requiring early diagnosis for better outcomes. While deep learning models have shown promise in dermatological image analysis, their performance is constrained by limited training data. Generative Adversarial Networks (GANs) offer a solution by generating synthetic images for data augmentation. However, assessing their clinical reliability remains difficult as automated metrics may not fully capture visual realism or clinical usability.The ObjectiveThis study presents a comprehensive framework for creating high-quality synthetic dermatoscopic pictures of MM lesions, as well as a holistic validation methodology that uses quantitative metrics and qualitative dermatologist assessment to provide a full clinical evaluation of the generated images. Three GAN architectures (DCGAN, StyleGAN2, and StyleGAN3-t) are explored. Lesions on the face, palms, and soles are excluded due to their unique dermoscopic patterns.Materials and MethodsA dataset of 1,774 dermatoscopic body MM images was used to train the models, assessed with Fréchet Inception Distance (FID), Kernel Inception Distance (KID), precision, and recall. Afterwards, a panel of 17 dermatologists with different levels of expertise assessed image quality using a 7-point Likert scale, with accuracy, sensitivity, specificity, and inter-rater agreement analysed.ResultsStyleGAN2 achieved the best image fidelity (FID: 18.89, KID: 0.0025), while StyleGAN3-t demonstrated stable but slower convergence. Both StyleGAN models outperformed DCGAN in diversity and fidelity. The validation study showed that StyleGAN2-generated images were often indistinguishable from real ones, reflected in low specificity and accuracy values among evaluators.ConclusionsThe study highlights the effectiveness of GANs in generating high-quality synthetic images, proposing a validation framework that integrates expert assessments with state-of-the-art quantitative metrics. This approach advances standardisation in GAN evaluation, ensuring synthetic images are clinically relevant for dermatological AI applications.

Generative adversarial networks (GANs) with gaming abilities have been widely applied in image generation. However, gamistic generators and discriminators may reduce the robustness of the obtained GANs in image generation under varying scenes. Enhancing the relation of hierarchical information in a generation network and enlarging differences of different network architectures can facilitate more structural information to improve the generation effect for image generation. In this paper, we propose an enhanced GAN via improving a generator for image generation (EIGGAN). EIGGAN applies a spatial attention to a generator to extract salient information to enhance the truthfulness of the generated images. Taking into relation the context account, parallel residual operations are fused into a generation network to extract more structural information from the different layers. Finally, a mixed loss function in a GAN is exploited to make a tradeoff between speed and accuracy to generate more realistic images. Experimental results show that the proposed method is superior to popular methods, i.e., Wasserstein GAN with gradient penalty (WGAN-GP) in terms of many indexes, i.e., Frechet Inception Distance, Learned Perceptual Image Patch Similarity, Multi-Scale Structural Similarity Index Measure, Kernel Inception Distance, Number of Statistically-Different Bins, Inception Score and some visual images for image generation.

Generative Adversarial Networks (GANs) are a major advancement in generative modeling, surpassing traditional machine learning models in tasks such as image generation, super-resolution, and image-to-text translation. A GAN consists of two neural networks: a Generator (G), which creates data from noise or a latent vector, and a Discriminator (D), which determines whether the data is real or generated. These networks train competitively, improving each other iteratively to produce increasingly realistic outputs. However, GANs face challenges like mode collapse, unstable training, and convergence issues, leading to the adoption of strategies such as instance normalization and enhanced loss functions. Future research can focus on improving stability, developing novel loss functions, and applying GANs in unsupervised learning. Performance metrics like Inception Score, Fréchet Inception Distance (FID), and Structural Similarity Index (SSIM) are essential for evaluating and comparing GAN architectures. Additionally, ethical concerns, including the misuse of GANs for deepfakes and synthetic data, underscore the importance of transparency, accountability, and ethical standards in research and deployment. This review provides an accessible introduction to GANs for novice researchers, along with a detailed analysis of their limitations, applications, and future prospects, offering valuable insights and direction for advancing this field.

We present a novel approach to developing robust AI in light of context-varying situations. This methodology harnesses a suite of indicators to establish a Robustness Assurance Quotient (RAQ) tailored to address environmentally noisy data while maintaining parity with current standards, namely the Fréchet Inception Distance (FID) metric. While the FID metric successfully indicates the closeness of highly structured data, it is less successful measuring closeness of inherently noisy data. A score for ascertaining the closeness between Generative Adversarial Network (GAN)-simulated, inherently noisy radiofrequnecy data and original GAN signals are proposed by focusing on the subject and context of the image. The RAQ demonstrably improved our GAN’s generated content similarity to original waterfall images of radiofrequency signal. We believe our Robustness Assurance Quotient can have a profound impact on improving the robustness of various AI models, in diverse application domains ultimately reducing the negative impact of noisy data.

AimsTo develop a generative adversarial network (GAN) capable of generating realistic high-resolution anterior segment optical coherence tomography (AS-OCT) images.MethodsThis study included 142 628 AS-OCT B-scans from the American University of...

Javanese script is an important part of Indonesia’s cultural heritage, representing cultural values from the past. However, recognizing and classifying Javanese characters within manuscripts is challenging due to the limited availability of data and uneven distribution of character classes. The decline in formal use of Javanese script has drastically reduced the pool of manuscript samples, causing certain characters to appear rarely and skewing class frequencies. Existing methods that utilize Generative Adversarial Networks (GANs) attempt to address this problem. However, they often struggle to generate characters that are both consistent and visually accurate in terms of structural details. To address these issues, this study introduces a skeleton-based balancing GAN (SkelBAGAN), which improves the structural details of the previous method for generating characters. The proposed method introduces three main enhancements: (i) a layer for extracting the character skeleton structure, (ii) an optimized pretrained network using an autoencoder for learning the skeleton distribution, and (iii) refinement of the evaluation function, preserving both the distribution and structural fidelity in the adversarial process. The performance of the proposed model is evaluated against previous methods using the Fréchet Inception Distance (FID) to assess distribution quality and the Structural Similarity Index Measure (SSIM) to evaluate structural fidelity. The results indicate that the proposed methods outperform previous methods in balancing the FID and SSIM metrics. The integration of all enhancements in SkelBAGAN achieves the lowest FID, indicating improved generative quality while maintaining competitive SSIM values. The qualitative study indicates that SkelBAGAN outperforms previous methods in character generation. These results highlight how the skeleton-based improvement of the quality of generated characters enhances the recognition performance for underrepresented Javanese characters in imbalanced datasets. Ultimately, this work contributes to the broader effort to preserve the Javanese script as a vital element of Indonesia’s cultural identity.

Generative Adversarial Networks (GANs) can be used for data augmentation in order to improve the outcome and performance of machine learning models for automatic information retrieval. We looked into the challenge faced with limited blurry and distorted digit images from expiry dates datasets, which is required to improve digit recognition tasks on medicine, consumables, cosmetic products and tube-type ointments. For our dataset, Wasserstein GAN with a gradient norm penalty (WGAN-GP) was effective for data augmentation among the state-of-the-art GANs by visible inspection and Fréchet Inception Distance (FID) value comparison.

Despite recent advances in digital banking around the world, the use of banknotes is still predominant. In this context, the detection of fake money bills is important due to the severe impact of their circulation on the economic system. Automating and making detecting counterfeited bills accessible to anyone is still an under-investigated issue, which is mainly based on the visual inspection of bills. In this paper, we propose a novel semi-supervised generative adversarial network (GAN)-based detection approach that, by visually inspecting a bill, can determine if the latter is authentic or fake. Our method originally combines a parallelized GAN to a semi-supervised GAN. On a dataset of Tunisian bills, our method is able to achieve 100% accuracy, with a Fréchet Inception Distance (FID) below 1, i.e., generating diversified fake images with good quality. These performances are proven superior to those of several benchmark approaches.

Generative Artificial Intelligence (AI) has shown significant promise in healthcare, particularly in medical imaging, by improving data augmentation, diagnostic accuracy, and model generalization. However, existing models such as Generative Adversarial Networks (GANs) and Diffusion Models face challenges in balancing image realism, structural fidelity, and computational efficiency. GANs often suffer from mode collapse and limited diversity, while diffusion models, though high in fidelity, are computationally intensive and slow to sample. To address these limitations, we propose a Hybrid GAN–Diffusion approach that synergistically combines the strengths of both architectures. Our pipeline first pretrains a StyleGAN2-ADA model to capture coarse-to-fine radiographic structure efficiently, then refines the generated samples using a Denoising Diffusion Probabilistic Model (DDPM) to enhance anatomical plausibility and diversity. This sequential design fuses the sharpness and speed of GANs with the coverage and robustness of diffusion models, yielding synthetic images of superior quality and clinical utility. We evaluate the framework on the Kaggle Chest X-Ray Pneumonia dataset for training and perform external validation on NIH ChestXray14 to assess cross-dataset generalization. Experimental results demonstrate that the hybrid model achieves an ROC-AUC of 0.984 in-distribution and 0.892 on the external dataset, outperforming classical augmentation, GAN-only, and diffusion-only baselines. The synthetic data quality is further validated by a simulated Fréchet Inception Distance (FID) of 16.2, indicating high realism and diversity. The proposed method establishes a new state-of-the-art in medical image augmentation, demonstrating that hybrid generative pipelines can effectively overcome data scarcity while improving model robustness and generalization.

In the process of intelligentization in modern manufacturing, especially in industrial fields such as automobile manufacturing, semiconductor production, and electronic product assembly, product quality control is crucial. Traditional defect detection methods face problems such as supervised learning methods relying on a large amount of labeled data, weak generalization ability, and high cost. In the process of intelligent manufacturing, industrial product quality control is a key link to ensure production safety and product consistency. Especially in typical industrial scenarios such as automobile manufacturing, semiconductor production, and electronic product assembly, traditional defect detection methods are difficult to meet the needs of actual production lines for high-precision and high-efficiency detection due to their reliance on a large amount of labeled data, weak generalization ability, and high cost. To solve the problem of defect detection in small and zero sample scenarios in these industries, improve detection sensitivity and localization accuracy, and enhance the model's generalization ability to unknown defects, a unsupervised industrial image defect detection method based on autoencoder and Generative Adversarial Networks (GANs) is proposed. This study constructs an Multi-level Deep feature Adaptive fusion AutoEncoder (MDAAE) module, extracts multi-scale features through Pre-Trained Convolutional Neural Backbone Networks (PTCNBN), introduces Attention Mechanism (AM) to dynamically calculate feature weights, and achieves feature fusion and reconstruction. Meanwhile, the self-AM is fused to improve the GANs. The self-attention module efficiently captures long-range dependencies and generates an adversarial network training objective function to optimize the generator's unsaturated loss. The results showed that the Area Under the Curve (AUC) of the Recall-Precision curve of the research method reached 93.6 ± 0.5%, and its F1-Score value exceeded 0.890 ± 0.003 in defect types such as scratches and dents. In practical applications, the inference time of the research method remained stable at 3.0 ± 0.2 GB of central processor memory, and the Fréchet Inception Distance (FID) value fully converged at the 2230th iteration, with a stable FID value of 3. The false alarm rate was only 8.6 ± 0.7% under strong light conditions. The proposed UIIDD method based on autoencoder and GANs had good robustness, generalization ability, reliability, and efficiency. This study effectively solves the problems of poor robustness, weak generalization ability, and high cost of traditional defect detection methods.

<p>High data collection costs and complicated data access regulations increase the demand for synthetic data. Generative Adversarial Networks (GANs) are a novel generative framework with great potential for high quality synthetic data generation. GANs formulate the true distribution of data implicitly, and the success of GANs are often measured based on the similarity of generated data to this true distribution. GANs were originally designed to work with continuous data. However, many important real-world datasets such as medical images involve discontinuous distributions. GAN training for discontinuous distributions is relatively more challenging, as the training procedure often suffers from instability and mode collapse issues. This dissertation focuses on designing novel GAN architectures to generate representative synthetic image data, and proposes new structures to alleviate GANs' mode collapse issue. As part of this thesis, novel applications of image data generation with GANs have been also investigated for important problems arising in the telecommunication industry and medical domain. Specifically, we first explore various GAN structures to generate engineered electromagnetic surfaces. We consider the continuous approximation of the data and explore the capabilities of feed-forward and convolutional GANs for synthetic data generation. Next, we introduce a novel GAN architecture to address the problem of mode collapse in GAN training. The proposed structure incorporates a third network that penalizes the generator for generating low diversity samples. Lastly, we study the challenging problem of object generation in 3D space using GANs, and we propose extensions to existing 3D GAN structures to generate connected 3D volumes. Additionally, we explore a more challenging version of this 3D volume generation problem by generating connected volumes packed with spheres. This research has applications in radiosurgery treatment planning, and the proposed 3D GAN structure can help generate rare, unseen 3D tumor volumes and information on how to treat these tumors. Accordingly, our analysis contributes to overcoming data scarcity issues (e.g., due to privacy considerations) for an important practical problem in the medical domain.</p>

This paper proposes a set of methods that enables low precision posit ™ arithmetic to be successfully used for the training of generative adversarial networks (GANs) with minimal quality loss. We show that ultra low precision posits, as small as 6 bits, can achieve high quality output for the generation phase after training. We also evaluate floating-point (float) formats and compare them to 8-bit posits in the context of GAN training. Our scaling and adaptive calibration techniques are capable of producing superior training quality for 8-bit posits that surpasses 8-bit floats and matches the results of 16-bit floats. Hardware simulation results indicate that our methods have higher energy efficiency compared to both 16- and 8-bit float training systems.

Generative AI has demonstrated strong capabilities in learning data distributions and producing realistic outputs. Although traditional approaches like Generative Adversarial Networks (GANs) have largely focused on static images and often fail to model coherent temporal sequences. Diffusion models have recently emerged as a more stable and effective alternative for generating time-sequenced data, yet their application to satellite imagery remains limited. Satellite data poses unique challenges such as multispectral channels and irregular temporal intervals that are poorly addressed by models trained on natural image datasets. To bridge this gap, we propose a spatio-temporal video diffusion model tailored for satellite-based forecasting tasks. Trained on curated datasets from the Landsat and Sentinel missions, our model generates temporally coherent sequences by conditioning on metadata like year, while effectively handling the spectral diversity and uneven sampling intervals characteristic of satellite imagery. When evaluated against retrained state-of-the-art baselines, our method demonstrates superior performance in modelling environmental changes, particularly deforestation, and achieves strong scores on perceptual quality metrics such as Fr´echet Inception Distance (FID) and Learned Perceptual Image Patch Similarity (LPIPS), highlighting its effectiveness for geospatial generative tasks. Our code and data is available at github.com/dfki-av/STDS.

The generation task from text to image generates cross modal data with consistent content by mining the semantic consistency contained in two different modal information of text and image. Due to the differences between the two modes, the task of text to image generation faces many difficulties and challenges. In this paper, we propose to boost the text-to-image synthesis through an adaptive learning and generating generative adversarial networks (ALG-GANs). First, we propose an adaptive forgetting mechanism in the generator to reduce the error accumulation and learn knowledge flexibly in the cascade structure. Besides, to evade the mode collapse caused by a strong biased surveillance, we propose a multi-task discriminator using weaksupervision information to guide the generator more comprehensively and maintain the semantic consistency in the cascade generation process. To avoid the refine difficulty aroused by the bad initialization, we judge the quality of initialization before further processing. The generator will re-sample the noise and re-initialize the bad initializations to obtain good ones. All the above contributions have been integrated in a unified framework, which is an adaptive forgetting, drafting and comprehensive guiding based text-to-image synthesis method with hierarchical generative adversarial networks. The model is evaluated on the Caltech-UCSD Birds 200 (CUB) dataset and the Oxford 102 Category Flowers (Oxford) dataset with standard metrics. The results on Inception Score (IS) and Fréchet Inception Distance (FID) show that our model outperforms the previous methods.