Adversarial Generative Network Research Articles

Vessel-promoted OCT to OCTA image translation by heuristic contextual constraints

Optical Coherence Tomography Angiography (OCTA) is a crucial tool in the clinical screening of retinal diseases, allowing for accurate 3D imaging of blood vessels through non-invasive scanning. However, the hardware-based approach for acquiring OCTA images presents challenges due to the need for specialized sensors and expensive devices. In this paper, we introduce a novel method called TransPro, which can translate the readily available 3D Optical Coherence Tomography (OCT) images into 3D OCTA images without requiring any additional hardware modifications. Our TransPro method is primarily driven by two novel ideas that have been overlooked by prior work. The first idea is derived from a critical observation that the OCTA projection map is generated by averaging pixel values from its corresponding B-scans along the Z-axis. Hence, we introduce a hybrid architecture incorporating a 3D adversarial generative network and a novel Heuristic Contextual Guidance (HCG) module, which effectively maintains the consistency of the generated OCTA images between 3D volumes and projection maps. The second idea is to improve the vessel quality in the translated OCTA projection maps. As a result, we propose a novel Vessel Promoted Guidance (VPG) module to enhance the attention of network on retinal vessels. Experimental results on two datasets demonstrate that our TransPro outperforms state-of-the-art approaches, with relative improvements around 11.4% in MAE, 2.7% in PSNR, 2% in SSIM, 40% in VDE, and 9.1% in VDC compared to the baseline method. The code is available at: https://github.com/ustlsh/TransPro.

Machine anomalous sound detection based on audio synthesis generative adversarial network

Abstract In the pursuit of implementing safety monitoring within Industry 4.0 smart production, deep learning methodologies are employed to harness the audio signals emanating from the operation of machinery and equipment, facilitating anomaly detection. Aiming to address the challenge of the few-shot samples for anomalous sound detection in an unsupervised setting, this paper introduces a machine anomalous sound detection method using an Audio Synthesis Adversarial Generative Network (ASGAN). The augmentation of classifiers within the network’s discriminator is proposed to execute a multi-classification task, thereby transforming the unsupervised training milieu into a semi-supervised one and enhancing the discriminator’s accuracy. To address the oversight of audio frequency domain features by conventional generative adversarial networks, the squeeze-and-excitation network is applied to concentrate on crucial frequency bands within the frequency domain. Experimental validation is conducted by using the DCASE Challenge Task2 dataset, revealing that the performance of the anomaly detection method leveraging audio synthesis generative adversarial networks surpasses that of the baseline system.

Identification of propagation path and root cause of faults based on generative adversarial networks in industrial systems

Fault diagnosis and detection (FDD) improves product quality, process safety, and profitability in modern industrial plants. Faults can propagate between different parts of plants due to connections between devices and control loops. As a result, control room operators may receive multiple alarms that distract their attention from addressing the main causes of abnormal situations. Therefore, investigating causal relationships between process variables is essential in determining the root cause of faults. This study proposes a new systematic approach for causality analysis to identify the propagation path and root cause of faults in industrial plants. The research's innovations are categorized into two main parts. Initially, the study utilizes machine learning algorithms to identify causal relationships using probabilistic concepts. Additionally, it introduces a novel architecture by merging the generative adversarial network (GAN) with the variational auto-encoder (VAE). The process variables are mapped to a latent space, where an adversarial generative network will investigate the causal relationship between variables. Finally, the Tennessee Eastman Process simulation and gas turbine industrial data are used to evaluate the proposed algorithm.

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

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

Empowering precise advertising with Fed-GANCC: A novel federated learning approach leveraging Generative Adversarial Networks and group clustering.

In the realm of targeted advertising, the demand for precision is paramount, and the traditional centralized machine learning paradigm fails to address this necessity effectively. Two critical challenges persist in the current advertising ecosystem: the data privacy concerns leading to isolated data islands and the complexity in handling non-Independent and Identically Distributed (non-IID) data and concept drift due to the specificity and diversity in user behavior data. Current federated learning frameworks struggle to overcome these hurdles satisfactorily. This paper introduces Fed-GANCC, an innovative federated learning framework that synergizes Generative Adversarial Networks (GANs) and Group Clustering. The framework incorporates a user data augmentation algorithm predicated on adversarial generative networks to enrich user behavior data, curtail the impact of non-uniform data distribution, and enhance the applicability of the global machine learning model. Unlike traditional approaches, our framework offers user data augmentation algorithms based on adversarial generative networks, which not only enriches user behavior data but also reduces the challenges posed by non-uniform data distribution, thereby enhancing the applicability of the global machine learning (ML) model. The effectiveness of Fed-GANCC is distinctly showcased through experimental results, outperforming contemporary methods like FED-AVG and FED-SGD in terms of accuracy, loss value, and receiver operating characteristic (ROC) indicators within the same computing time. Experimental results vindicate the effectiveness of Fed-GANCC, revealing substantial enhancements in accuracy, loss value, and receiver operating characteristic (ROC) metrics compared to FED-AVG and FED-SGD given the same computational time. These outcomes underline Fed-GANCC's exceptional prowess in mitigating issues such as isolated data islands, non-IID data, and concept drift. With its novel approach to addressing the prevailing challenges in targeted advertising such as isolated data islands, non-IID data, and concept drift, the Fed-GANCC framework stands as a benchmark, paving the way for future advancements in federated learning solutions tailored for the advertising domain. The Fed-GANCC framework promises to offer pivotal insights for the future development of efficient and advanced federated learning solutions for targeted advertising.

Maximum Classifier Discrepancy Generative Adversarial Network for Jointly Harmonizing Scanner Effects and Improving Reproducibility of Downstream Tasks.

Multi-site collaboration is essential for overcoming small-sample problems when exploring reproducible biomarkers in MRI studies. However, various scanner-specific factors dramatically reduce the cross-scanner replicability. Moreover, existing harmony methods mostly could not guarantee the improved performance of downstream tasks. we proposed a new multi-scanner harmony framework, called 'maximum classifier discrepancy generative adversarial network', or MCD-GAN, for removing scanner effects in the original feature space while preserving substantial biological information for downstream tasks. Specifically, the adversarial generative network was utilized for persisting the structural layout of each sample, and the maximum classifier discrepancy module was introduced for regulating GAN generators by incorporating the downstream tasks. We compared the MCD-GAN with other state-of-the-art data harmony approaches (e.g., ComBat, CycleGAN) on simulated data and the Adolescent Brain Cognitive Development (ABCD) dataset. Results demonstrate that MCD-GAN outperformed other approaches in improving cross-scanner classification performance while preserving the anatomical layout of the original images. To the best of our knowledge, the proposed MCD-GAN is the first generative model which incorporates downstream tasks while harmonizing, and is a promising solution for facilitating cross-site reproducibility in various tasks such as classification and regression.

A Comprehensive Brain MRI Image Segmentation System Based on Contourlet Transform and Deep Neural Networks

Brain tumors are one of the deadliest types of cancer. Rapid and accurate identification of brain tumors, followed by appropriate surgical intervention or chemotherapy, increases the probability of survival. Accurate determination of brain tumors in MRI scans determines the exact location of surgical intervention or chemotherapy. However, this accurate segmentation of brain tumors, due to their diverse morphologies in MRI scans, poses challenges that require significant expertise and accuracy in image interpretation. Despite significant advances in this field, there are several barriers to proper data collection, particularly in the medical sciences, due to concerns about the confidentiality of patient information. However, research papers for learning systems and proposed networks often rely on standardized datasets because a specific approach is unavailable. This system combines unsupervised learning in the adversarial generative network component with supervised learning in segmentation networks. The system is fully automated and can be applied to tumor segmentation on various datasets, including those with sparse data. In order to improve the learning process, the brain MRI segmentation network is trained using a generative adversarial network to increase the number of images. The U-Net model was employed during the segmentation step to combine the remaining blocks efficiently. Contourlet transform produces the ground truth for each MRI image obtained from the adversarial generator network and the original images in the processing and mask preparation phase. On the part of the adversarial generator network, high-quality images are produced, the results of which are similar to the histogram of the original images. Finally, this system improves the image segmentation performance by combining the remaining blocks with the U-net network. Segmentation is evaluated using brain magnetic resonance images obtained from Istanbul Medipol Hospital. The results show that the proposed method and image segmentation network, which incorporates several criteria, such as the DICE criterion of 0.9434, can be effectively used in any dataset as a fully automatic system for segmenting different brain MRI images.

A Physics-informed Deep Neural Network for Harmonization of CT Images.

An adversarial generative network was trained on virtual CT images acquired under various imaging conditions using a virtual imaging platform with 40 computational patient models. These models featured anthropomorphic lungs with different levels of pulmonary diseases, including nodules and emphysema. Imaging was conducted using a validated CT simulator at two dose levels and varying reconstruction kernels. The trained model was tested on an independent virtual test dataset and two clinical datasets. On the virtual test set, the harmonizer improved the structural similarity index from 79.3 ±16.4% to 95.8 ±1.7%, normalized mean squared error from 16.7 ±9.7% to 9.2 ±1.7%, and peak signal-to-noise ratio from 27.7 ±3.7dB to 32.2 ±1.6dB. Moreover, the harmonized images yielded more precise quantification of emphysema-based imaging biomarkers for lung attenuation, LAA-950 from 5.6 ±8.7% to 0.23 ±0.16%, Perc15 from 43.4 ±45.4 HU to 20.0 ±7.5 HU, and Lung Mass from 0.3 ±0.3 g to 0.1 ±0.2 g. In clinical data, the harmonizer reduced biomarker variability by an average of 70%. For lung nodules, harmonized images improved the detectability index by 6.5-fold and DNN-based precision by 6%. The proposed harmonizer significantly enhances image quality and quantification accuracy in CT imaging. The study demonstrated the potential utility of image harmonization for consistent CT image quality and reliable quantification, which is crucial for clinical applications and patient management.

Insulator Defect Recognition Based on Vision Big‐Model Transfer Learning and Stochastic Configuration Network

Insulator faults are an important factor in causing outages and accidents in power transmission lines. In response to problems related to inefficient insulator positioning, limited robustness of insulator defect feature extraction methods, and the scarcity of defective insulator samples leading to poor classifier generalization, a method for insulator defect detection and recognition based on vision big‐model transfer learning and a stochastic configuration network (SCN) is proposed. First, data augmentation methods, such as Mosaic and Mixup, are employed to mitigate overfitting in the YOLOv7 network. Second, StyleGanv3 adversarial generative networks are used to augment the dataset of defective insulators, which enhances dataset diversity. Third, a vision big‐model transfer learning method based on DINOv2 is introduced to extract features from insulator images. Finally, an SCN classifier is used to determine the status of insulators. Experimental results demonstrate that the applied data augmentation methods effectively mitigate overfitting. YOLOv7 accurately detects insulator positions, and the use of the DINOv2 feature extraction method increases the accuracy of insulator defect recognition by 28.6%. Compared with machine learning classification methods, the SCN classifier achieves the highest accuracy improvement of 17.4%. The proposed method effectively detects insulator positions and recognizes insulator defects.

Realistic image generation using adversarial generative networks combined with depth information

Existing image generation tasks produce blurry, unrealistic results and images that lack layers and structure. Depth information can be used to accurately control the relative positions and hierarchies between different objects in an image. Our goal is to enhance the realism, hierarchy, and quality of generated images by using depth information in image-to-image tasks. To address these issues, we propose a multi-conditional semantic image generation method that fuses depth information. The method is based on the network structure of Generative Adversarial Networks and fuses the depth information of multi-conditional inputs by using pairs of semantic labels and depth maps as inputs through our proposed Multi-scale Feature Extraction and Information Fusion Module. Furthermore, we add a channel-attention mechanism to the generator to strengthen the interconnectivity between channels and suppress confusion between different semantic features. With less increase in training cost, the module proposed in this paper can generate real images that match the input semantic layout. Through extensive testing on three challenging datasets, the images generated by this model produce superior visuals and data metrics, demonstrating the effectiveness of our proposed method.

A 2D image 3D reconstruction function adaptive denoising algorithm.

To address the issue of image denoising algorithms blurring image details during the denoising process, we propose an adaptive denoising algorithm for the 3D reconstruction of 2D images. This algorithm takes into account the inherent visual characteristics of human eyes and divides the image into regions based on the entropy value of each region. The background region is subject to threshold denoising, while the target region undergoes processing using an adversarial generative network. This network effectively handles 2D target images with noise and generates a 3D model of the target. The proposed algorithm aims to enhance the noise immunity of 2D images during the 3D reconstruction process and ensure that the constructed 3D target model better preserves the original image's detailed information. Through experimental testing on 2D images and real pedestrian videos contaminated with noise, our algorithm demonstrates stable preservation of image details. The reconstruction effect is evaluated in terms of noise reduction and the fidelity of the 3D model to the original target. The results show an average noise reduction exceeding 95% while effectively retaining most of the target's feature information in the original image. In summary, our proposed adaptive denoising algorithm improves the 3D reconstruction process by preserving image details that are often compromised by conventional denoising techniques. This has significant implications for enhancing image quality and maintaining target information fidelity in 3D models, providing a promising approach for addressing the challenges associated with noise reduction in 2D images during 3D reconstruction.

Deep Learning for Detecting Multi-Level Driver Fatigue Using Physiological Signals: A Comprehensive Approach.

A large share of traffic accidents is related to driver fatigue. In recent years, many studies have been organized in order to diagnose and warn drivers. In this research, a new approach was presented in order to detect multi-level driver fatigue. A multi-level driver tiredness diagnostic database based on physiological signals including ECG, EEG, EMG, and respiratory effort was developed for this aim. The EEG signal was used for processing and other recorded signals were used to confirm the driver's fatigue so that fatigue was not confirmed based on self-report questionnaires. A customized architecture based on adversarial generative networks and convolutional neural networks (end-to-end) was utilized to select/extract features and classify different levels of fatigue. In the customized architecture, with the objective of eliminating uncertainty, type 2 fuzzy sets were used instead of activation functions such as Relu and Leaky Relu, and the performance of each was investigated. The final accuracy obtained in the three scenarios considered, two-level, three-level, and five-level, were 96.8%, 95.1%, and 89.1%, respectively. Given the suggested model's optimal performance, which can identify five various levels of driver fatigue with high accuracy, it can be employed in practical applications of driver fatigue to warn drivers.

The application of big data technology in the art of visual communication design with information technology

Abstract Based on big data technology, this paper firstly proposes that visual image generation design based on the improved adversarial generative network becomes one of the ideas to solve the current dilemma of visual communication design, which is not only a convenient, creative means but also changes the traditional way of visual communication design. Secondly, in the extension of visual image samples, in addition to using some operations such as rotation, translation and scale transformation of visual images, an improved DCGAN adversarial generative network algorithm is also used. The improved DCGAN network is then used to augment the number of visual images while maintaining the quality of visual image generation. Finally, intelligent generation models based on visual communication design processes and thinking are investigated, focusing on intelligent data sets and retrieval, image generation techniques, and intelligent solution recommendations. The results show that DCGAN has a lower value of IS = 1.5, FID = 249.39, IS and a higher value of FID, indicating that it generates better quality visual images and greatly improves the efficiency and accuracy of visual communication design. This study can extract effective information and generate visual images more accurately and efficiently, and provide technical support for visual communication designers.

Quantitative assessment of adaptive radiotherapy for prostate cancer using deep learning: Bladder dose as a decision criterion.

Adaptive radiotherapy (ART) can incorporate anatomical variations in a reoptimized treatment plan for fractionated radiotherapy. An automatic solution to objectively determine whether ART should be performed immediately after the daily image acquisition is highly desirable. We investigate a quantitative criterion for whether ART should be performed in prostate cancer radiotherapy by synthesizing pseudo-CT (sCT) images and evaluating dosimetric impact on treatment planning using deep learning approaches. Planning CT (pCT) and daily cone-beam CT (CBCT) data sets of 74 patients are used to train (60 patients) and evaluate (14 patients) a cycle adversarial generative network (CycleGAN) that performs the task of synthesizing high-quality sCT from daily CBCT. Automatic delineation (AD) of the bladder is performed on the sCT using the U-net. The combination of sCT and AD allows us to perform dose calculations based on the up-to-date bladder anatomy to determine whether the original treatment plan (ori-plan) is still applicable. For positive cases that the patients' anatomical changes and the associated dose calculations warrant re-planning, we made rapid plan revisions (re-plan) based on the ori-plan. The mean absolute error within the region-of-interests (i.e., body, bladder, fat, muscle) between the sCT and pCT are 41.2, 25.1, 26.5, and 29.0HU, respectively. Taking the calculated results of pCT doses as the standard, for PTV, the gamma passing rates of sCT doses at 1mm/1%, 2mm/2% are 87.92%, 98.78%, respectively. The Dice coefficients of the AD-contours are 0.93 on pCT and 0.91 on sCT. According to the result of dose calculation, we found when the bladder volume underwent a substantial change (79.7%), the bladder dose is still within the safe limit, suggesting it is insufficient to solely use the bladder volume change as a criterion to determine whether adaptive treatment needs to be done. After AD-contours of the bladder using sCT, there are two cases whose bladder dose . For the two cases, we perform re-planning to reduce the bladder dose to , under the condition that the PTV meets the prescribed dose. We provide a dose accurate adaptive workflow for prostate cancer patients by using deep learning approaches, and implement ART that adapts to bladder dose. Of note, the specific replanning criterion for whether ART needs to be performed can adapt to different centers' choices based on their experience and daily observations.

EMA‐GAN: A Generative Adversarial Network for Infrared and Visible Image Fusion with Multiscale Attention Network and Expectation Maximization Algorithm

The purpose of the infrared and visible image fusion is to generate a fused image with rich information. Although most fusion methods can achieve good performance, there are still shortcomings in extracting feature information from source images, which make it difficult to balance the thermal radiation region information and texture detail information in the fused image. To address the above issues, an expectation maximization (EM) learning framework based on adversarial generative networks (GAN) for infrared and visible image fusion is proposed. The EM algorithm (EMA) can obtain maximum likelihood estimation for problems with potential variables, which is helpful in solving the problem of lack of labels in infrared and visible image fusion. The axial‐corner attention mechanism is designed to capture long‐range semantic information and texture information of the visible image. The multifrequency attention mechanism digs the relationships between features at different scales to highlight target information of infrared images in the fused result. Meanwhile, two discriminators are used to balance two different features, and a new loss function is designed to maximize the likelihood estimate of the data with soft class label assignments, which is obtained from the expectation network. Extensive experiments demonstrate the superiority of EMA‐GAN over the state‐of‐the‐art.

DEGAN: Decompose-Enhance-GAN Network for Simultaneous Low-Light Image Lightening and Denoising

Images taken in low-light situations frequently have a significant quality reduction. Taking care of these degradation problems in low-light images is essential for raising their visual quality and enhancing high-level visual task performance. However, because of the inherent information loss in dark images, conventional Retinex-based approaches for low-light image enhancement frequently fail to accomplish real denoising. This research introduces DEGANet, a revolutionary deep-learning framework created particularly for improving and denoising low-light images. To overcome these restrictions, DEGANet makes use of the strength of a Generative Adversarial Network (GAN). The Decom-Net, Enhance-Net, and an Adversarial Generative Network (GAN) are three linked subnets that make up our novel Retinex-based DEGANet architecture. The Decom-Net is in charge of separating the reflectance and illumination components from the input low-light image. This decomposition enables Enhance-Net to effectively enhance the illumination component, thereby improving the overall image quality. Due to the complicated noise patterns, fluctuating intensities, and intrinsic information loss in low-light images, denoising them presents a significant challenge. By incorporating a GAN into our architecture, DEGANet is able to effectively denoise and smooth the enhanced image as well as retrieve the original data and fill in the gaps, producing an output that is aesthetically beautiful while maintaining key features. Through a comprehensive set of studies, we demonstrate that DEGANet exceeds current state-of-the-art methods in both terms of image enhancement and denoising quality.

Pretraining technology of adversarial generative networks based data selection

The effectiveness of adversarial generative networks (GANs) has been shown in many works. However, the convergence of the GANs should be granted mathematically with proper hyperparameters and some network structures. Considering the difficulties of constructing a large and clean dataset, GANs are required in many cases to extract enough information from a relatively small dataset. This could result in bad coverage performance of GANs, and one way to tackle this problem is pretraining. Since this method relies on importing external information from the pretraining dataset, this paper does a necessary discussion on the effect that the pretraining dataset has on the model and makes a simple assumption on the standard of choosing a pretraining dataset. Specifically, this paper focuses on the case where the target is to generate animation-style characters, and it has been proven that when providing a pretraining dataset with higher diversity, the performance of the GAN is better. The result of this paper should reveal the possibility that using a specially chosen pretraining dataset can improve the result of GAN.

MS-ACGAN: A modified auxiliary classifier generative adversarial network for schizophrenia's samples augmentation based on microarray gene expression data

Artificial intelligence-based models and robust computational methods have expedited the data-to-knowledge trajectory in precision medicine. Although machine learning models have been widely applied in medical data analysis, some barriers are yet to be challenging, such as available biosample shortage, prohibitive costs, rare diseases, and ethical considerations. Transcriptomics, an omics approach that studies gene activities and provides gene expression data such as microarray and RNA-Sequences faces the difficulties of biospecimen collection, particularly for mental disorders, as some psychiatric patients avoid medical care. Microarray data suffers from the low number of available samples, making it challenging to apply machine learning models. However, adversarial generative network (GAN), the hottest paradigm in deep learning, has created unprecedented momentum in data augmentation and efficiently expands datasets. This paper proposes a novel model termed MS-ACGAN, where the generator feeds on a bordered Gaussian distribution. In machine learning, calibration is of utmost importance, which gives insight into model uncertainty and is considered a crucial step toward improving the robustness and reliability of models. Therefore, we apply calibration techniques to classifiers and focus on estimating their probabilities as accurately as possible. Additionally, we present our trustworthy outputs by harnessing confidence intervals that confine the point estimate limitations and report a range of expected values for performance metrics. Both concepts statistically describe the implemented model's reliability in this study. Furthermore, we employ two quantitative measures, GAN-train and GAN-test, to demonstrate that the artificial data generated by our robust approach remarkably resembles the original data characteristics.

Driver Decision Space Inversion Method Based on DC-GAN

In the driving process, drivers need to constantly perceive the surrounding environment to make decisions and perform operations. This cognitive space formed in the driver's mind is the driver's decision space. For the actual driving environment, factors such as complex road sign settings, unreasonable road planning, long-time fatigue driving, and reduced reaction ability of elderly drivers may interfere with the normal perception of the driver's decision space, resulting in reduced driving safety. In this paper, a driver decision space inversion method based on deep convolutional neural networks and generative adversarial networks is proposed to study driver perception in near-domain traffic scenarios. The steps of the method include: near-domain target element extraction, driving data collection, sample data generation, adversarial generative network model learning, data enhancement and decision space inversion. The experimental results show that the method in this paper can accurately identify the driver's decision space in both real and simulated driving scenarios by implementing real-time monitoring of driver perception. The method is important for studying the driver's decision space, which can promote the development of intelligent driving technology and the synergistic development of human-vehicle-road-loop. In the context of today's sustainable transportation and smart mobility, driver decision space research is not only an important basic research, but also an inevitable requirement to promote innovation and upgrading in the transportation field.

Bubble detection in photoresist with small samples based on GAN augmentations and modified YOLO

Photoresist coating is a key procedure in semiconductor wafer surface processing, and bubbles in a photoresist dropper usually affect the uniformity of photoresist film, which seriously reduces the etching quality of semiconductor wafer surface. However, the problems of blurred edges of photoresist bubbles and complex corrugations in photoresist itself make the automatic detection of photoresist bubbles very challenging. To overcome this problem, a high-precision photoresist bubble detection method based on deep learning is proposed in this paper. First, aiming at the problem that the photoresist bubble samples are extremely lacking and cannot meet the requirement of deep learning model training, this paper proposes a sample automatic generator bubbleGAN based on the adversarial generative network, so as to effectively increase the number of photoresist bubble samples with better diversity of the target to be detected, thus improving the performance of the detection model; Second, aiming at the challenges of target detection caused by the blurred edges of photoresist bubbles and the complex ripples of photoresist itself, this paper proposes a novel bubble detection method mYOLO. Based on the recently developed high-performance target detection method YOLOX, we optimize its structure and activation function to reduce the difficulty of model training, solve the dead zone problem of activation function, and improve the accuracy of model detection. The experimental results on the dataset collected by the actual rotary coating machine prove the effectiveness of the proposed algorithm. In addition, we further compare the impact of different manual data augmentation methods on the detection model.