Frechet Inception Distance Research Articles

EEG Data Augmentation Method for Identity Recognition Based on Spatial–Temporal Generating Adversarial Network

Traditional identity recognition methods are facing significant security challenges due to their vulnerability to leakage and forgery. Brainprint recognition, a novel biometric identification technology leveraging EEG signals, has emerged as a promising alternative owing to its advantages such as resistance to coercion, non-forgeability, and revocability. Nevertheless, the scarcity of high-quality electroencephalogram (EEG) data limits the performance of brainprint recognition systems, necessitating the use of shallow models that may not perform optimally in real-world scenarios. Data augmentation has been demonstrated as an effective solution to address this issue. However, EEG data encompass diverse features, including temporal, frequency, and spatial components, posing a crucial challenge in preserving these features during augmentation. This paper proposes an end-to-end EEG data augmentation method based on a spatial–temporal generative adversarial network (STGAN) framework. Within the discriminator, a temporal feature encoder and a spatial feature encoder were parallelly devised. These encoders effectively captured global dependencies across channels and time of EEG data, respectively, leveraging a self-attention mechanism. This approach enhances the data generation capabilities of the GAN, thereby improving the quality and diversity of the augmented EEG data. The identity recognition experiments were conducted on the BCI-IV2A dataset, and Fréchet inception distance (FID) was employed to evaluate data quality. The proposed method was validated across three deep learning models: EEGNET, ShallowConvNet, and DeepConvNet. Experimental results indicated that data generated by STGAN outperform DCGAN and RGAN in terms of data quality, and the identity recognition accuracies on the three networks were improved by 2.49%, 2.59% and 1.14%, respectively.

Style harmonization of panoramic radiography using deep learning.

This study aimed to harmonize panoramic radiograph images from different equipment in a single institution to display similar styles. A total of 15,624 panoramic images were acquired using two different equipment: 8079 images from Rayscan Alpha Plus (R-unit) and 7545 images from Pax-i plus (P-unit). Among these, 222 image pairs (444 images) from the same patients comprised the test dataset to harmonize the P-unit images with the R-unit image style using CycleGAN. Objective evaluations included Frechet Inception Distance (FID) and Learned Perceptual Image Patch Similarity (LPIPS) assessments. Additionally, expert evaluation was conducted by two oral and maxillofacial radiologists on transformed P-unit and R-unit images. The statistical analysis of LPIPS employed a Student's t-test. The FID and mean LPIPS values of the transformed P-unit images (7.362, 0.488) were lower than those of the original P-unit images (8.380, 0.519), with a significant difference in LPIPS (p < 0.05). The experts evaluated 43.3-46.7% of the transformed P-unit images as R-unit images, 20.0-28.3% as P-units, and 28.3-33.3% as undetermined images. CycleGAN has the potential to harmonize panoramic radiograph image styles. Enhancement of the model is anticipated for the application of images produced by additional units.

Facial expression morphing: enhancing visual fidelity and preserving facial details in CycleGAN-based expression synthesis

Recent advancements in facial expression synthesis using deep learning, particularly with Cycle-Consistent Adversarial Networks (CycleGAN), have led to impressive results. However, a critical challenge persists: the generated expressions often lack the sharpness and fine details of the original face, such as freckles, moles, or birthmarks. To address this issue, we introduce the Facial Expression Morphing (FEM) algorithm, a novel post-processing method designed to enhance the visual fidelity of CycleGAN-based outputs. The FEM method blends the input image with the generated expression, prioritizing the preservation of crucial facial details. We experimented with our method on the Radboud Faces Database (RafD) and evaluated employing the Fréchet Inception Distance (FID) standard benchmark for image-to-image translation and introducing a new metric, FSD (Facial Similarity Distance), to specifically measure the similarity between translated and real images. Our comprehensive analysis of CycleGAN, UNet Vision Transformer cycle-consistent GAN versions 1 (UVCGANv1) and 2 (UVCGANv2) reveals a substantial enhancement in image clarity and preservation of intricate details. The average FID score of 31.92 achieved by our models represents a remarkable 50% reduction compared to the previous state-of-the-art model’s score of 63.82, showcasing the significant advancements made in this domain. This substantial enhancement in image quality is further supported by our proposed FSD metric, which shows a closer resemblance between FEM-processed images and the original faces.

Advancing Electric Engineering Education through Immersive Virtual Reality: Deep Learning and Evolutionary Algorithms for Image Stitching and Rectification in Virtual Lab Environments

Virtual Reality (VR) technology has emerged as a transformative tool in education, offering immersive and interactive experiences that enhance learning outcomes. This paper delves into the application of image stitching and rectification techniques to create a VR lab environment, specifically tailored for electrical engineering education. The importance of VR technology in education is explored, highlighting its role in promoting active learning and providing experiential learning opportunities. The primary emphasis of this Paper lies in the smooth incorporation of image stitching algorithms for the creation of panoramic perspectives, along with the implementation of rectification techniques to correct irregular borders within the stitched images. By utilizing Convolutional Neural Networks (CNNs) and Genetic Algorithms (GAs), the proposed approach optimizes the rectification process, resulting in visually cohesive representations. Demonstrating the utilization of the VR lab across a range of situations, such as examining power transfer and creating control panels for water pumps in irrigation initiatives, the immersive setting enables students to delve into intricate systems. The performance of the proposed method was evaluated using various metrics, including mean squared error, peak signal to noise ratio (PSNR), structural similarity index (SSIM), and Fréchet inception distance (FID). the combination of deep learning algorithm specifically (CNN) and optimization algorithm specifically (Genetic algorithm (GA)) led to an increase in the accuracy of the rectified images where the average PSNR reached 23.98, SSIM was 0.8066, and FID was 18.72. Regarding the users’ opinion about the generated environment by stitching and rectifying images, participants demonstrated consistent positive sentiments, with mean scores ranging from 3.65 to 4.03, all above the scale midpoint, and moderate variability indicated by standard deviation values ranging from 1.070 to 1.251, suggesting general favorability with some variation in responses. This experience empowers the users to gain insights and cultivate essential problemsolving abilities at a heightened level. Collaborative learning is facilitated, enabling students to engage in collaborative projects regardless of their physical location. Through the synthesis of image processing techniques and VR technology, this research contributes to the enrichment of educational experiences and the advancement of electrical engineering education.

A novel 8-connected Pixel Identity GAN with Neutrosophic (ECP-IGANN) for missing imputation

Missing pixel imputation presents a critical challenge in image processing and computer vision, particularly in applications such as image restoration and inpainting. The primary objective of this paper is to accurately estimate and reconstruct missing pixel values to restore complete visual information. This paper introduces a novel model called the Enhanced Connected Pixel Identity GAN with Neutrosophic (ECP-IGANN), which is designed to address two fundamental issues inherent in existing GAN architectures for missing pixel generation: (1) mode collapse, which leads to a lack of diversity in generated pixels, and (2) the preservation of pixel integrity within the reconstructed images. ECP-IGANN incorporates two key innovations to improve missing pixel imputation. First, an identity block is integrated into the generation process to facilitate the retention of existing pixel values and ensure consistency. Second, the model calculates the values of the 8-connected neighbouring pixels around each missing pixel, thereby enhancing the coherence and integrity of the imputed pixels. The efficacy of ECP-IGANN was rigorously evaluated through extensive experimentation across five diverse datasets: BigGAN-ImageNet, the 2024 Medical Imaging Challenge Dataset, the Autonomous Vehicles Dataset, the 2024 Satellite Imagery Dataset, and the Fashion and Apparel Dataset 2024. These experiments assessed the model’s performance in terms of diversity, pixel imputation accuracy, and mode collapse mitigation, with results demonstrating significant improvements in the Inception Score (IS) and Fréchet Inception Distance (FID). ECP-IGANN markedly enhanced image segmentation performance in the validation phase across all datasets. Key metrics, such as Dice Score, Accuracy, Precision, and Recall, were improved substantially for various segmentation models, including Spatial Attention U-Net, Dense U-Net, and Residual Attention U-Net. For example, in the 2024 Medical Imaging Challenge Dataset, the Residual Attention U-Net’s Dice Score increased from 0.84 to 0.90, while accuracy improved from 0.88 to 0.93 following the application of ECP-IGANN. Similar performance enhancements were observed with the other datasets, highlighting the model’s robust generalizability across diverse imaging domains.

AI-ADC: Channel and Spatial Attention-Based Contrastive Learning to Generate ADC Maps from T2W MRI for Prostate Cancer Detection.

Apparent Diffusion Coefficient (ADC) maps in prostate MRI can reveal tumor characteristics, but their accuracy can be compromised by artifacts related with patient motion or rectal gas associated distortions. To address these challenges, we propose a novel approach that utilizes a Generative Adversarial Network to synthesize ADC maps from T2-weighted magnetic resonance images (T2W MRI). By leveraging contrastive learning, our model accurately maps axial T2W MRI to ADC maps within the cropped region of the prostate organ boundary, capturing subtle variations and intricate structural details by learning similar and dissimilar pairs from two imaging modalities. We trained our model on a comprehensive dataset of unpaired T2-weighted images and ADC maps from 506 patients. In evaluating our model, named AI-ADC, we compared it against three state-of-the-art methods: CycleGAN, CUT, and StyTr2. Our model demonstrated a higher mean Structural Similarity Index (SSIM) of 0.863 on a test dataset of 3240 2D MRI slices from 195 patients, compared to values of 0.855, 0.797, and 0.824 for CycleGAN, CUT, and StyTr2, respectively. Similarly, our model achieved a significantly lower Fréchet Inception Distance (FID) value of 31.992, compared to values of 43.458, 179.983, and 58.784 for the other three models, indicating its superior performance in generating ADC maps. Furthermore, we evaluated our model on 147 patients from the publicly available ProstateX dataset, where it demonstrated a higher SSIM of 0.647 and a lower FID of 113.876 compared to the other three models. These results highlight the efficacy of our proposed model in generating ADC maps from T2W MRI, showcasing its potential for enhancing clinical diagnostics and radiological workflows.

A comprehensive investigation on the performance of reconstruction of noncircular fiber-representative volume elements in unidirectional composites using diffusion generative models

This study employs diffusion generative models to reconstruct random representative volume elements (RVEs) of unidirectional composites with noncircular fibers. Microscope images of these composites were prepared and trained with denoising diffusion probabilistic model (DDPM), denoising diffusion implicit model (DDIM), progressive distillation diffusion model (PDDM), and deep convolutional generative adversarial network (DCGAN). Hyperparameter tuning was performed for both DDPM and PDDM, and the generated RVE images were evaluated using the two-point correlation function (TPCF), Fréchet Inception Distance (FID), and computational cost. Furthermore, finite element (FE) models were generated using these images, and FE simulations were conducted considering interfacial debonding behavior. The resulting stress and strain curves from these simulations were compared. The results show that DDPM demonstrated the best performance in final image quality, while PDDM maintained stable performance from the early stages of training. Additionally, both models exhibited excellent agreement with the original images, indicating high quality, diversity, and resemblance.

Conditional probabilistic diffusion model driven synthetic radiogenomic applications in breast cancer.

This study addresses the heterogeneity of Breast Cancer (BC) by employing a Conditional Probabilistic Diffusion Model (CPDM) to synthesize Magnetic Resonance Images (MRIs) based on multi-omic data, including gene expression, copy number variation, and DNA methylation. The lack of paired medical images and genomics data in previous studies presented a challenge, which the CPDM aims to overcome. The well-trained CPDM successfully generated synthetic MRIs for 726 TCGA-BRCA patients, who lacked actual MRIs, using their multi-omic profiles. Evaluation metrics such as Frechet's Inception Distance (FID), Mean Square Error (MSE), and Structural Similarity Index Measure (SSIM) demonstrated the CPDM's effectiveness, with an FID of 2.02, an MSE of 0.02, and an SSIM of 0.59 based on the 15-fold cross-validation. The synthetic MRIs were used to predict clinical attributes, achieving an Area Under the Receiver-Operating-Characteristic curve (AUROC) of 0.82 and an Area Under the Precision-Recall Curve (AUPRC) of 0.84 for predicting ER+/HER2+ subtypes. Additionally, the MRIs served to accurately predicted BC patient survival with a Concordance-index (C-index) score of 0.88, outperforming other baseline models. This research demonstrates the potential of CPDMs in generating MRIs based on BC patients' genomic profiles, offering valuable insights for radiogenomic research and advancements in precision medicine. The study provides a novel approach to understanding BC heterogeneity for early detection and personalized treatment.

Faceify- Human Face Sketch to Real

The Human Face Sketch to Real Image Application using deep learning presents the development of an innovative application designed to convert human face sketches into realistic image using advanced deep learning techniques. The primary objective of this research is to bridge the gap between manual sketching and digital realism, offering a powerful tool for artists, law enforcement, and digital content creators. The proposed system leverages generative adversarial networks (GANs), particularly a variant known as the Sketch-to-Image GAN (SI-GAN), which is optimizes for translating sparse and abstract sketch lines into high-fidelity facial representations. The architecture consists of a generator network that refines sketches into realistic images, and a discriminator network that distinguishes between generated images and real photographs, continuously enhancing the model’s output. To ensure high accuracy and detail preservation, the model is trained on a diverse dataset comprising thousands of paired sketch and real face images. Advanced techniques such as attention mechanisms and multi-scale feature extraction are incorporated to maintain the integrity of facial features and textures. Additionally, a user friendly interface has been developed, allowing users to input sketches and receive high-quality realistic images in real-time. Evaluation metrics, including structural similarity index (SSIM) and Frechet Inception Distance (FID), Indicate that the proposed model outperforms existing state-of-the-art methods in both quality and realism. Keywords: Image Translation, Conditional Generative Adversarial Networks (CGANs), Deep Learning, Mobile Application Development, Backend Integration, Face Sketch, Photorealistic Image.

Generating Synthesized Fluorescein Angiography Images From Color Fundus Images by Generative Adversarial Networks for Macular Edema Assessment.

To assess the feasibility of generating synthetic fluorescein angiography (FA) images from color fundus (CF) images using pixel-to-pixel generative adversarial network (pix2pixGANs) for clinical applications. Research questions addressed image realism to retinal specialists and utility for assessing macular edema (ME) in Retinal Vein Occlusion (RVO) eyes. We used a registration-guided pix2pixGANs method trained on the CF-FA dataset from Kham Eye Centre, Kandze Prefecture People's Hospital. A visual Turing test confirmed the realism of synthetic images without novel artifacts. We then assessed the synthetic FA images for assessing ME. Finally, we quantitatively evaluated the synthetic images using Fréchet Inception distance (FID) and structural similarity measures (SSIM). The raw development dataset had 881 image pairs from 349 subjects. Our approach is capable of generating realistic FA images because small vessels are clearly visible and sharp within one optic disc diameter around the macula. Two retinal specialists agreed that more than 85% of synthetic FA images have good or excellent image quality. For ME detection, accuracy was similar for real and synthetic images. FID demonstrated a 38.9% improvement over the previous state-of-the-art (SOTA), and SSIM reached 0.78 compared to the previous SOTA's 0.67. We developed a pix2pixGANs model translating FA images from label-free CF images, yielding reliable synthetic FA images. This suggests potential for noninvasive evaluation of ME in RVO eyes using pix2pix GANs techniques. Pix2pixGANs techniques have the potential to assist in the noninvasive clinical assessment of ME in RVO eyes.

Generative AI in orthopedics: an explainable deep few-shot image augmentation pipeline for plain knee radiographs and Kellgren-Lawrence grading.

Recently, deep learning medical image analysis in orthopedics has become highly active. However, progress has been restricted by the absence of large-scale and standardized ground-truth images. To the best of our knowledge, this study is the first to propose an innovative solution, namely a deep few-shot image augmentation pipeline, that addresses this challenge by synthetically generating knee radiographs for training downstream tasks, with a specific focus on knee osteoarthritis Kellgren-Lawrence (KL) grading. This study leverages a deep few-shot image augmentation pipeline to generate synthetic knee radiographs. Despite the limited availability of training samples, we demonstrate the capability of our proposed computational strategy to produce high-fidelity plain knee radiographs and use them to successfully train a KL grade classifier. Our experimental results showcase the effectiveness of the proposed computational pipeline. The generated synthetic radiographs exhibit remarkable fidelity, evidenced by the achieved average Frechet Inception Distance (FID) score of 26.33 for KL grading and 22.538 for bilateral knee radiographs. For KL grading classification, the classifier achieved a test Cohen's Kappa and accuracy of 0.451 and 0.727, respectively. Our computational strategy also resulted in a publicly and freely available imaging dataset of 86000 synthetic knee radiographs. Our approach demonstrates the capability to produce top-notch synthetic knee radiographs and use them for KL grading classification, even when working with a constrained training dataset. The results obtained emphasize the effectiveness of the pipeline in augmenting datasets for knee osteoarthritis research, opening doors for broader applications in orthopedics, medical image analysis, and AI-powered diagnosis.

Method for Augmenting Side-Scan Sonar Seafloor Sediment Image Dataset Based on BCEL1-CBAM-INGAN.

In this paper, a method for augmenting samples of side-scan sonar seafloor sediment images based on CBAM-BCEL1-INGAN is proposed, aiming to address the difficulties in acquiring and labeling datasets, as well as the insufficient diversity and quantity of data samples. Firstly, a Convolutional Block Attention Module (CBAM) is integrated into the residual blocks of the INGAN generator to enhance the learning of specific attributes and improve the quality of the generated images. Secondly, a BCEL1 loss function (combining binary cross-entropy and L1 loss functions) is introduced into the discriminator, enabling it to focus on both global image consistency and finer distinctions for better generation results. Finally, augmented samples are input into an AlexNet classifier to verify their authenticity. Experimental results demonstrate the excellent performance of the method in generating images of coarse sand, gravel, and bedrock, as evidenced by significant improvements in the Frechet Inception Distance (FID) and Inception Score (IS). The introduction of the CBAM and BCEL1 loss function notably enhances the quality and details of the generated images. Moreover, classification experiments using the AlexNet classifier show an increase in the recognition rate from 90.5% using only INGAN-generated images of bedrock to 97.3% using images augmented using our method, marking a 6.8% improvement. Additionally, the classification accuracy of bedrock-type matrices is improved by 5.2% when images enhanced using the method presented in this paper are added to the training set, which is 2.7% higher than that of the simple method amplification. This validates the effectiveness of our method in the task of generating seafloor sediment images, partially alleviating the scarcity of side-scan sonar seafloor sediment image data.

Creating realistic anterior segment optical coherence tomography images using generative adversarial networks

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...

Performance evaluation of generative adversarial networks for anime face synthesis using deep learning approaches

Anime characters have transcended their traditional animation role, permeating entertainment, video games, and social media. Online platforms, including prominent blogs and forums, are filled with references and adaptations of these characters. This digital ubiquity has spurred creators to develop innovative storytelling methods through the internet, fostering new avenues for anime character creation. Notably, advanced systems like Generative Adversarial Networks (GANs) have emerged as powerful tools for this purpose. This research investigates the capabilities of various GAN variants in generating anime faces. The Fréchet Inception Distance (FID) serves as a key metric for performance comparison. The analysis focuses on Deep Convolutional GAN (DCGAN), Progressively Growing GAN (ProGAN), and Style GAN2. Additionally, the performance metrics of the DCGAN model are examined. The findings show demonstrate significant differences in FID scores across these GANs. Notably, Style GAN2 exhibits superior performance in generating anime faces, achieving a significantly lower FID score (31) compared to DCGAN (625) and ProGAN (218.3). This research underscores Style GAN2's superior effectiveness in creating anime characters within the digital landscape.

Clinical performance of deep learning-enhanced ultrafast whole-body scintigraphy in patients with suspected malignancy

BackgroundTo evaluate the clinical performance of two deep learning methods, one utilizing real clinical pairs and the other utilizing simulated datasets, in enhancing image quality for two-dimensional (2D) fast whole-body scintigraphy (WBS).MethodsA total of 83 patients with suspected bone metastasis were retrospectively enrolled. All patients underwent single-photon emission computed tomography (SPECT) WBS at speeds of 20 cm/min (1x), 40 cm/min (2x), and 60 cm/min (3x). Two deep learning models were developed to generate high-quality images from real and simulated fast scans, designated 2x-real and 3x-real (images from real fast data) and 2x-simu and 3x-simu (images from simulated fast data), respectively. A 5-point Likert scale was used to evaluate the image quality of each acquisition. Accuracy, sensitivity, specificity, and the area under the curve (AUC) were used to evaluate diagnostic efficacy. Learned perceptual image patch similarity (LPIPS) and the Fréchet inception distance (FID) were used to assess image quality. Additionally, the count-level consistency of WBS was compared between the two models.ResultsSubjective assessments revealed that the 1x images had the highest general image quality (Likert score: 4.40 ± 0.45). The 2x-real, 2x-simu and 3x-real, 3x-simu images demonstrated significantly better quality than the 2x and 3x images (Likert scores: 3.46 ± 0.47, 3.79 ± 0.55 vs. 2.92 ± 0.41, P < 0.0001; 2.69 ± 0.40, 2.61 ± 0.41 vs. 1.36 ± 0.51, P < 0.0001), respectively. Notably, the quality of the 2x-real images was inferior to that of the 2x-simu images (Likert scores: 3.46 ± 0.47 vs. 3.79 ± 0.55, P = 0.001). The diagnostic efficacy for the 2x-real and 2x-simu images was indistinguishable from that of the 1x images (accuracy: 81.2%, 80.7% vs. 84.3%; sensitivity: 77.27%, 77.27% vs. 87.18%; specificity: 87.18%, 84.63% vs. 87.18%. All P > 0.05), whereas the diagnostic efficacy for the 3x-real and 3x-simu was better than that for the 3x images (accuracy: 65.1%, 66.35% vs. 59.0%; sensitivity: 63.64%, 63.64% vs. 64.71%; specificity: 66.67%, 69.23% vs. 55.1%. All P < 0.05). Objectively, both the real and simulated models achieved significantly enhanced image quality from the accelerated scans in the 2x and 3x groups (FID: 0.15 ± 0.18, 0.18 ± 0.18 vs. 0.47 ± 0.34; 0.19 ± 0.23, 0.20 ± 0.22 vs. 0.98 ± 0.59. LPIPS: 0.17 ± 0.05, 0.16 ± 0.04 vs. 0.19 ± 0.05; 0.18 ± 0.05, 0.19 ± 0.05 vs. 0.23 ± 0.04. All P < 0.05). The count-level consistency with the 1x images was excellent for all four sets of model-generated images (P < 0.0001).ConclusionsUltrafast 2x speed (real and simulated) images achieved comparable diagnostic value to that of standardly acquired images, but the simulation algorithm does not necessarily reflect real data.

Bone metastasis scintigram generation using generative adversarial learning with multi-receptive field learning and two-stage training.

Deep learning is the primary method for conducting automated analysis of SPECT bone scintigrams. The lack of available large-scale data significantly hinders the development of well-performing deep learning models, as the performance of a deep learning model is positively correlated with the size of the dataset used. Therefore, there is an urgent demand for an automated data generation method to enlarge the dataset of SPECT bone scintigrams. We introduce a deep learning-based generation model that can generate realistic but not identical samples from the original SPECT bone scintigrams. Following the generative adversarial learning architecture, a bone metastasis scintigram generation model christened BMS-Gen is proposed. First, BMS-Gen takes multiple input conditions and employs multi-receptive field learning to ensure that the generated samples are as realistic as possible. Second, BMS-Gen adopts generative adversarial learning to retain the diversity of the generated samples. Last, BMS-Gen uses a two-stage training strategy to improve the quality of the generated samples. Experimental evaluation conducted on a set of clinical data of SPECT BM scintigrams has shown the performance of the proposed BMS-Gen, achieving the best overall scores of 1678.0, 69.33, and 19.51 for FID (Fréchet Inception Distance), MSE (Mean Square Error), and PSNR (Peak Signal-to-Noise Ratio) metrics. The introduction of samples generated by BMS-Gen contributes a maximum (minimum) increase of 3.01% (0.15%) on the F-1 score and a maximum (minimum) increase of 6.83% (2.21%) on the DSC score for the image classification and segmentation tasks, respectively. The proposed BMS-Gen model can be used as a promising tool for augmenting the data of bone scintigrams, greatly facilitating the development of deep learning-based automated analysis of SPECT bone scintigrams.

An attempt to generate panoramic radiographs including jaw cysts using StyleGAN3.

The purpose of this study was to generate radiographs including dentigerous cysts by applying the latest generative adversarial network (GAN; StyleGAN3) to panoramic radiography. A total of 459 cystic lesions were selected, and 409 images were randomly assigned as training data and 50 images as test data. StyleGAN3 training was performed for 500 000 images. Fifty generated images were objectively evaluated by comparing them with 50 real images according to four metrics: Fréchet inception distance (FID), kernel inception distance (KID), precision and recall, and inception score (IS). A subjective evaluation of the generated images was performed by three specialists who compared them with the real images in a visual Turing test. The results of the metrics were as follows: FID, 199.28; KID, 0.14; precision, 0.0047; recall, 0.00; and IS, 2.48. The overall results of the visual Turing test were 82.3%. No significant difference was found in the human scoring of root resorption. The images generated by StyleGAN3 were of such high quality that specialists could not distinguish them from the real images.

A Broad Generative Network for Two-Stage Image Outpainting.

Image outpainting is a challenge for image processing since it needs to produce a big scenery image from a few patches. In general, two-stage frameworks are utilized to unpack complex tasks and complete them step-by-step. However, the time consumption caused by training two networks will hinder the method from adequately optimizing the parameters of networks with limited iterations. In this article, a broad generative network (BG-Net) for two-stage image outpainting is proposed. As a reconstruction network in the first stage, it can be quickly trained by utilizing ridge regression optimization. In the second stage, a seam line discriminator (SLD) is designed for transition smoothing, which greatly improves the quality of images. Compared with state-of-the-art image outpainting methods, the experimental results on the Wiki-Art and Place365 datasets show that the proposed method achieves the best results under evaluation metrics: the Fréchet inception distance (FID) and the kernel inception distance (KID). The proposed BG-Net has good reconstructive ability with faster training speed than those of deep learning-based networks. It reduces the overall training duration of the two-stage framework to the same level as the one-stage framework. Furthermore, the proposed method is adapted to image recurrent outpainting, demonstrating the powerful associative drawing capability of the model.

Local fusion generative adversarial network with dual-discriminator and parallel multipath and its application in machinery fault diagnosis with imbalanced data

Abstract Diagnosing faults in critical machinery components is imperative for effective condition monitoring and real-world datasets often suffer from data imbalance. To address this issue, numerous data generation methods have been developed, such as improved local fusion generative adversarial network (ILoFGAN), variational autoencoding GAN (VAEGAN), etc. However, the existing data generation methods primarily concentrate on global and single-scale features and often ignore local or multi-scale features, which leads to the omission of key features or nuances in the generated data. Therefore, a novel approach called the Local Fusion Generative Adversarial Network with Dual-discriminator and Parallel multipath (LoFGAN-DP) is designed to enhance the fault diagnosis performance in the context of imbalanced data. The LoFGAN-DP features a Parallel Multi-Path (PMP) module along with a dual-discriminator scheme, in which the multipath module facilitates feature extraction at various scales through convolution across paths of diverse sizes, and the dual-discriminator scheme can better improve the quality and diversity of the samples generated by the generator. The PMP module and dual-discriminator scheme enhance the proposed method's robustness against variations in input data. After generating data by LoFGAN-DP, a two-dimensional capsule network is further used to achieve the efficient recognition of fault features. To validate the proposed LoFGAN-DP in the machinery fault diagnosis with imbalanced data, the gear dataset and the self-constructed bearing dataset were utilized. Experimental results show that LoFGAN-DP significantly improves Structural Similarity Index (SSIM), Fréchet Inception Distance (FID), and fault classification accuracy compared to several advanced methods.

Adversarial enhancement generation method for side-scan sonar images based on DDPM–YOLO

As marine exploration advances, efficient and accurate seabed target identification becomes increasingly critical. However, traditional methods and current technologies face challenges such as scarce samples and complex imaging conditions when dealing with side-scan sonar images. Given the scarcity of sample augmentation methods for side-scan sonar, this paper iteratively trains the Denoising Diffusion Probabilistic Models(DDPM), integrating the DDPM diffusion model and the downstream You Only Look Once(YOLO) retrieval task into a mutually reinforcing framework, proposing an adversarial enhancement generation method based on the DDPM and YOLO detection models. Experiments demonstrate that the DDPM model generated through this adversarial enhancement generation method can improve the accuracy of downstream YOLO target detection tasks by 7%. The images generated by this model also perform optimally on the Fréchet Inception Distance(FID), Maximum Mean Discrepancy(MMD), and Learned Perceptual Image Patch Similarity(LPIPS) metrics, thereby proving that our method can enhance the quality of generated images from the side-scan sonar diffusion model and offer a new avenue for improving the construction of underwater target detection models.