Sketch-guided spatial adaptive normalization and high-level feature constraints based GAN image synthesis for steel strip defect detection data augmentation

A major challenge in applying deep learning to medical imaging is the paucity of annotated data. This study explores the use of synthetic images for data augmentation to address the challenge of limited annotated data in colonoscopy lesion classification. We demonstrate that synthetic colonoscopy images generated by Generative Adversarial Network (GAN) inversion can be used as training data to improve polyp classification performance by deep learning models. We invert pairs of images with the same label to a semantically rich and disentangled latent space and manipulate latent representations to produce new synthetic images. These synthetic images maintain the same label as the input pairs. We perform image modality translation (style transfer) between white light and narrow-band imaging (NBI). We also generate realistic synthetic lesion images by interpolating between original training images to increase the variety of lesion shapes in the training dataset. Our experiments show that GAN inversion can produce multiple colonoscopy data augmentations that improve the downstream polyp classification performance by 2.7% in F1-score and 4.9% in sensitivity over other methods, including state-of-the-art data augmentation. Testing on unseen out-of-domain data also showcased an improvement of 2.9% in F1-score and 2.7% in sensitivity. This approach outperforms other colonoscopy data augmentation techniques and does not require re-training multiple generative models. It also effectively uses information from diverse public datasets, even those not specifically designed for the targeted downstream task, resulting in strong domain generalizability. Project code and model: https://github.com/DurrLab/GAN-Inversion.

AI-generated content (AIGC) in the context of dermoscopic image analysis describes the application of artificial intelligence (AI) approaches to produce synthetic images for training and enriching machine learning (ML) methods. In dermatology, where dermoscopic images are essential in skin lesion detection, AIGC helps overcome the challenges of inadequate datasets and class imbalances. By employing advanced generative complex systems like Generative Adversarial Networks (GANs), AIGC will simulate different dermoscopic conditions by generating realistic-view synthetic images. These AI-generated images help to enhance the training dataset, contributing to the ML models with many widespread and varied sets of instances. This aids in increasing the ability of the model to precisely predict and generalize while handling new, unnoticed dermoscopic images. Deep learning (DL) methods, specifically convolutional neural networks (CNNs), can provide a massive breakthrough in an extensive range of computer vision (CV) tasks of complex systems, predominantly by exploiting large-scale annotated datasets. This paper develops a GAN-based Synthetic Medical Image Augmentation for the imbalanced Dermoscopic Image Analysis (GANSMIA-CDIA) method. The GANSMIA-CDIA technique’s primary target is exploiting the GAN model for synthetic image generation to handle class imbalance data problems. In addition, the GANSMIA-CDIA technique diagnoses the melanoma using the optimal DL model. The GANSMIA-CDIA technique applies a contrast enhancement process for the noise eradication process. The GANSMIA-CDIA technique follows a feature fusion method comprising different DL methods such as MobileNetV2, AlexNet, ResNet50, and Inceptionv3 to learn feature patterns in the preprocessed images. Meanwhile, the Social Network Search (SNS) technique is utilized for the hyperparameter tuning process. Also, the bidirectional long short-term memory (BiLSTM) technique is implemented to detect and classify melanoma. A series of simulations were performed on the standard dermoscopic image dataset to evaluate the performance of the GANSMIA-CDIA technique. The experimental values indicate the excellence of the GANSMIA-CDIA technique over existing techniques.

The problem of fault classification in industry has been studied extensively. Most classification algorithms are modeled on the premise of data balance. However, the difficulty of collecting industrial data in different modes is quite different. This inevitably leads to data imbalance, which will adversely affect the fault classification performance. This article proposes a novel data augmentation classifier (DAC) for imbalanced fault classification. Data augmentation based on generative adversarial networks (GANs) is an effective way to solve the problem of unbalanced classification. However, the randomness of the GAN generation process restricts the effect of data enhancement. DAC proposes a data selection strategy based on data filtering and data purification in model training to solve this problem. In addition, DAC combines supervised learning and data generation processes to obtain an end-to-end model. Meanwhile, multigenerator structure of DAC (MDAC) is proposed to solve the problem of incomplete learning of a single generator when data imbalances get complicated. The proposed DAC and MDAC are applied in two fault classification cases of the Tennessee Eastman (TE) benchmark process, results of which show superiority of DAC and MDAC compared to existing methods. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Data imbalances are common in fault classification and affect the effectiveness of modeling in industry. As a generative model, generative adversarial networks (GANs) provide new ideas for small-class data augmentation. However, the instability of its training process and the randomness of data generation affect the results of data augmentation. In this article, the GAN generation process is analyzed in detail. The results of the visualization indicate that no data generation was perfect at any one time. Based on the rules of GAN data generation, we propose a data selection strategy during training. High-quality data are selected for data augmentation through data filtering and data purification. Apart from this, we combine the training process of GAN and classification model for imbalanced data to reduce modeling time. Through industrial examples, we have evaluated the effectiveness of this method.

This study aims to explore the feasibility of employing generative adversarial networks (GAN) to generate synthetic contrast-enhanced ultrasound (CEUS) from grayscale ultrasound images of patients with thyroid nodules while dispensing with the need for ultrasound contrast agents injection. Patients who underwent preoperative thyroid CEUS examinations between January 2020 and July 2022 were collected retrospectively. The cycle-GAN framework integrated paired and unpaired learning modules was employed to develop the non-invasive image generation process. The syntheticCEUS imageswas generated in three phases: pre-arterial, plateau, and venous. The evaluation included quantitative similarity metrics, classification performance, and qualitative assessment by radiologists. CEUS videos of 360 thyroid nodules from 314 patients (45 years ± 12 [SD]; 272 women) in the internal dataset and 202 thyroid nodules from 183 patients (46 years ± 13 [SD]; 148 women) in the external dataset were included. In the external testing dataset, quantitative analysis revealed a significant degree of similarity between real and synthetic CEUS images (structure similarity index, 0.89 ± 0.04; peak signal-to-noise ratio, 28.17 ± 2.42). Radiologists deemed 126 of 132 [95%] synthetic CEUS images diagnostically useful. The accuracy of radiologists in distinguishing between real and synthetic images was 55.6% (95% CI: 0.49, 0.63), with an AUC of 61.0% (95% CI: 0.65, 0.68). No statistically significant difference (p > 0.05) was observed when radiologists assessed peak intensity and enhancement patterns using real CEUS and synthetic CEUS. Both quantitative analysis and radiologist evaluations exhibited that synthetic CEUS images generated by generative adversarial networks were similar to real CEUS images. QuestionIt is feasible to generate synthetic thyroid contrast-enhanced ultrasound images using generative adversarial networks without ultrasound contrast agents injection. FindingsCompared to real contrast-enhanced ultrasound images, synthetic contrast-enhanced ultrasound images exhibited high similarity and image quality. Clinical relevanceThis non-invasive and intelligent transformation may reduce the requirement for ultrasound contrast agents in certain cases, particularly in scenarios where ultrasound contrast agents administration is contraindicated, such as in patients with allergies, poor tolerance, or limited access to resources.

Mammography is the primary procedure for breast cancer screening, attempting to reduce breast cancer mortality risk with early detection. Deep learning methods have shown strong applicability to various medical images datasets. Due to paucity of available labeled medical images, accurate computer assisted diagnosis requires intensive data augmentation (DA) techniques, such as geometric/intensity transformations of original images. This data when used along with the training data helps to address the limited medical image dataset collected from various sources. Generative Adversarial Networks (GANs) is one of the DA techniques. GAN trained on images can generate new images that contain many authentic characteristics and look realistic to human observers. Therefore, this paper focuses on overcoming the problem of limited labeled dataset, using Deep Convolution GANs (DCGANs). To analyze the closeness between the original and synthetic images, a visualization tool ImageJ was used. In order to validate the proposed model, a visual Turing test was conducted with the help of medical experts.

PurposeTo develop generative adversarial networks (GANs) that synthesize realistic anterior segment optical coherence tomography (AS-OCT) images and evaluate deep learning (DL) models that are trained on real and synthetic datasets for detecting angle closure.MethodsThe GAN architecture was adopted and trained on the dataset with AS-OCT images collected from the Joint Shantou International Eye Center of Shantou University and the Chinese University of Hong Kong, synthesizing open- and closed-angle AS-OCT images. A visual Turing test with two glaucoma specialists was performed to assess the image quality of real and synthetic images. DL models, trained on either real or synthetic datasets, were developed. Using the clinicians’ grading of the AS-OCT images as the reference standard, we compared the diagnostic performance of open-angle vs. closed-angle detection of DL models and the AS-OCT parameter, defined as a trabecular-iris space area 750 µm anterior to the scleral spur (TISA750), in a small independent validation dataset.ResultsThe GAN training included 28,643 AS-OCT anterior chamber angle (ACA) images. The real and synthetic datasets for DL model training have an equal distribution of open- and closed-angle images (all with 10,000 images each). The independent validation dataset included 238 open-angle and 243 closed-angle AS-OCT ACA images. The image quality of real versus synthetic AS-OCT images was similar, as assessed by the two glaucoma specialists, except for the scleral spur visibility. For the independent validation dataset, both DL models achieved higher areas under the curve compared with TISA750. Two DL models had areas under the curve of 0.97 (95% confidence interval, 0.96–0.99) and 0.94 (95% confidence interval, 0.92–0.96).ConclusionsThe GAN synthetic AS-OCT images appeared to be of good quality, according to the glaucoma specialists. The DL models, trained on all-synthetic AS-OCT images, can achieve high diagnostic performance.Translational RelevanceThe GANs can generate realistic AS-OCT images, which can also be used to train DL models.

Large annotated datasets are required for training deep learning models, but in medical imaging data sharing is often complicated due to ethics, anonymization and data protection legislation. Generative AI models, such as generative adversarial networks (GANs) and diffusion models, can today produce very realistic synthetic images, and can potentially facilitate data sharing. However, in order to share synthetic medical images it must first be demonstrated that they can be used for training different networks with acceptable performance. Here, we therefore comprehensively evaluate four GANs (progressive GAN, StyleGAN 1–3) and a diffusion model for the task of brain tumor segmentation (using two segmentation networks, U-Net and a Swin transformer). Our results show that segmentation networks trained on synthetic images reach Dice scores that are 80%–90% of Dice scores when training with real images, but that memorization of the training images can be a problem for diffusion models if the original dataset is too small. Our conclusion is that sharing synthetic medical images is a viable option to sharing real images, but that further work is required. The trained generative models and the generated synthetic images are shared on AIDA data hub.

Deep Learning algorithms have almost become a key standard for majority of vision and machine learning problems. Despite its common usage and high performance for many applications, they have certain disadvantages. One major problem with deep learning methods is the size of the dataset to be used for training. The methods require a large dataset for an adequate training. However, a large dataset may not be available for all problems. In such a case, researchers apply data augmentation methods to obtain a larger dataset from a given dataset. For the image classification problem, conventional method for data augmentation is the application of transformation based methods; such as flipping, rotation, blurring etc. Recently, generative models which apply deep learning methods are also commonly used for data augmentation. On the other hand, in case of a too large dataset the classifiers may overfit and end up with a lack of generalization. In this study, we explore the usage of generative adversarial networks for data augmentation in the image classification problem. We evaluate the classification performance with three types of augmentation methods. Dataset is first augmented by two conventional methods; Gaussian blurring and dropout of regions, then by generative adversarial networks. Meanwhile, we observe the behavior of the classifier for various sized datasets with and without data augmentation. We observe that especially in datasets of certain sizes generative adversarial networks can be effectively used for data augmentation.

The proliferation of Artificial Intelligence (AI) models such as Generative Adversarial Networks (GANs) has shown impressive success in image synthesis. Artificial GAN-based synthesized images have been widely spread over the Internet with the advancement in generating naturalistic and photo-realistic images. This might have the ability to improve content and media; however, it also constitutes a threat with regard to legitimacy, authenticity, and security. Moreover, implementing an automated system that is able to detect and recognize GAN-generated images is significant for image synthesis models as an evaluation tool, regardless of the input modality. To this end, we propose a framework for reliably detecting AI-generated images from real ones through Convolutional Neural Networks (CNNs). First, GAN-generated images were collected based on different tasks and different architectures to help with the generalization. Then, transfer learning was applied. Finally, several Class Activation Maps (CAM) were integrated to determine the discriminative regions that guided the classification model in its decision. Our approach achieved 100% on our dataset, i.e., Real or Synthetic Images (RSI), and a superior performance on other datasets and configurations in terms of its accuracy. Hence, it can be used as an evaluation tool in image generation. Our best detector was a pre-trained EfficientNetB4 fine-tuned on our dataset with a batch size of 64 and an initial learning rate of 0.001 for 20 epochs. Adam was used as an optimizer, and learning rate reduction along with data augmentation were incorporated.

Deep learning has gained importance in surface defect detection and has drawn significant research attention. However, sufficient defect samples with reliable labels are usually not achievable in real production, leading to inaccuracies in defect detection. Generating images with generative adversarial networks (GANs) may encounter issues of mode collapse or overfitting under limited data. A clustering-guided GAN (CG-GAN) with data augmentation is proposed to produce diverse, high-quality samples, expand defect datasets, and improve detection performance. The introduction of scene prediction, based on image clustering, can effectively label defective samples and reduce interference and confusion among defects. Additionally, it helps lower the cost of manual labeling. Furthermore, the defect clusters guide the discriminator to establish a multi-type defect decision surface for the fusion of classification and discrimination. The feedback from the discriminator, in the form of scene factors, is employed as multi-task noise input to the generator, thereby generating better-quality images and avoiding overfitting. Extensive experiments show that the proposed method enhances pre-trained DL networks to result in accuracy improvement of 1.56% and 0.5% on the Northeastern University surface defect public dataset, as well as 3.12% and 4.68% on the real printed circuit board (PCB) dataset, respectively. Thus, the proposed method effectively augments insufficient and imbalanced datasets, improving defect detection capabilities.

Image dehazing has always been a challenging topic in image processing. The development of deep learning methods, especially the generative adversarial networks (GAN), provides a new way for image dehazing. In recent years, many deep learning methods based on GAN have been applied to image dehazing. However, GAN has two problems in image dehazing. Firstly, For haze image, haze not only reduces the quality of the image but also blurs the details of the image. For GAN network, it is difficult for the generator to restore the details of the whole image while removing the haze. Secondly, GAN model is defined as a minimax problem, which weakens the loss function. It is difficult to distinguish whether GAN is making progress in the training process. Therefore, we propose a guided generative adversarial dehazing network (GGADN). Different from other generation adversarial networks, GGADN adds a guided module on the generator. The guided module verifies the network of each layer of the generator. At the same time, the details of the map generated by each layer are strengthened. Network training is based on the pre-trained VGG feature model and L1-regularized gradient prior which is developed by new loss function parameters. From the dehazing results of synthetic images and real images, the proposed method is better than the state-of-the-art dehazing methods.

In part due to its ability to mimic any data distribution, Generative Adversarial Network (GAN) algorithms have been successfully applied to many applications, such as data augmentation, text‐to‐image translation, image‐to‐image translation, and image inpainting. Learning from data without crafting loss functions for each application provides broader applicability of the GAN algorithm. Medical image synthesis is also another field that the GAN algorithm has great potential to assist clinician training. This paper proposes a synthetic wound image generation model based on GAN architecture to increase the quality of clinical training. The proposed model is trained on chronic wound datasets with various sizes taken from real hospital environments. Hyperparameters such as epoch count and dataset size for training tasks are studied to find optimum training conditions as well. The performance of the developed model was evaluated through a mean squared error (MSE) metric to determine the similarity between generated and actual wounds. Visual inspection is performed to examine generated wound images. The results show that the proposed synthetic wound image generation (WG2AN) model has great potential to be used in medical training and performs well in producing synthetic wound images with a 1000‐image training dataset and 200 epochs of training.

Deep learning-based synthetic contrast imaging has been proposed as an alternative to iodinated and gadolinium-based contrast agents in CT and MRI. This scoping review aimed to provide a cross-modality overview of current evidence, including study characteristics and validation strategies. Following PRISMA-ScR guidelines, PubMed, Embase, Scopus, and Web of Science were searched from inception to September 2025. Eligible studies applied deep learning to synthesize contrast-enhanced CT/MRI from non-contrast or modified-contrast inputs and reported reference-based validation. Extracted variables included modality, anatomy, input type, model type, dataset scale, validation category, and four predefined evaluation metrics. Fifty-six studies met the inclusion criteria (25 CT, 31 MRI). The brain was the most frequent target, followed by head and neck, breast, and liver applications. Non-contrast inputs were used in 71% of studies, with the remainder using modified-contrast strategies. Generative adversarial networks were the predominant model class, while diffusion and transformer models appeared after 2023. Dataset sizes ranged from 10 to 7306 (median, 218), and 57% of studies were single-center. Quantitative fidelity was evaluated in 88% of studies, reporting structural similarity index values of 0.73-0.99 and peak signal-to-noise ratios of 22-51 dB. Task-based performance was assessed in 39% of studies, radiologist-rated image quality in 54%, and diagnostic performance in 30%, with sensitivities of 72-92% and specificities of 59-95%. Deep learning-based synthetic contrast imaging shows high quantitative and perceptual fidelity, but evidence supporting diagnostic interchangeability and routine clinical use remains limited. Question Reducing contrast media use in CT and MRI is a clinical priority, and deep learning-based synthetic imaging is emerging as a potential alternative. Findings Deep learning generated synthetic contrast-enhanced images with high quantitative fidelity and acceptable image quality, but diagnostic performance was evaluated in only limited studies. Clinical relevance This review outlines the capabilities and limitations of deep learning-based synthetic contrast imaging, informing the development of contrast-minimizing strategies for safer CT and MRI practice.

Generative Adversarial Network (GAN) is an algorithmic architecture containing two neural networks, placed against each other to generate new synthetic images and it has been used successfully in image segmentation. The paper analyzes different GAN implementations for segmentation of images acquired by aerial robots in a real context of a rural zone in Romania. To improve the segmentation performance, a new GAN network is proposed by adding a new layer. Data augmentation was done by the following techniques: mirroring, rotation, scaling, gray scaling, blurring, sharpening, etc. Five classes of region of interest are considered: floods, vegetations, buildings, roads, and dry land. GAN implementations were tested on CPU, GPU, and TPU, on individual computing devices and in the cloud. A new layer was added. The performances were analyzed in terms of learning time, operating time, and statistical indicators. The batch size was generally low: batches of 1, 4 or 16 images were used in this paper. The results confirm that the use of batch achieves the best training and generalization performance in terms of computational cost, for a wide range of experiments.

This review paper examines the current state of lung disease diagnosis based on deep learning (DL) methods. Lung diseases, such as Pneumonia, TB, Covid-19, and lung cancer, are significant causes of morbidity and mortality worldwide. Accurate and timely diagnosis of these diseases is essential for effective treatment and improved patient outcomes. DL methods, which utilize artificial neural networks to extract features from medical images automatically, have shown great promise in improving the accuracy and efficiency of lung disease diagnosis. This review discusses the various DL methods that have been developed for lung disease diagnosis, including convolutional neural networks (CNNs), deep neural networks (DNNs), and generative adversarial networks (GANs). The advantages and limitations of each method are discussed, along with the types of medical imaging techniques used, such as X-ray and computed tomography (CT). In addition, the review discusses the most commonly used performance metrics for evaluating the performance of DL for lung disease diagnosis: the area under the curve (AUC), sensitivity, specificity, F1-score, accuracy, precision, and the receiver operator characteristic curve (ROC). Moreover, the challenges and limitations of using DL for lung disease diagnosis, including the limited availability of annotated data, the variability in imaging techniques and disease presentation, and the interpretability and generalizability of DL models, are highlighted in this paper. Furthermore, strategies to overcome these challenges, such as transfer learning, data augmentation, and explainable AI, are also discussed. The review concludes with a call for further research to address the remaining challenges and realize DL's full potential for improving lung disease diagnosis and treatment.