Data Augmentation Algorithms Research Articles

IRDA: Implicit data augmentation for deep imbalanced regression

Imbalanced data distributions are prevalent in real-world classification and regression tasks. Data augmentation is a commonly employed technique to mitigate this issue, with implicit methods gaining attention for their effectiveness and efficiency. However, implicit data augmentation methods have not been extensively explored in the context of regression tasks. To address this gap, we introduce IRDA, a novel learning method for regression that incorporates implicit data augmentation. Our approach includes developing a new augmentation strategy specifically tailored for deep imbalanced regression tasks, and a regression loss function that is suitable for real-world data with imbalanced label distributions and non-uniformly distributed features. We derive an easily computable surrogate loss and propose two implicit data augmentation algorithms, one incorporating meta-learning and one without. Additionally, we provide a regularization perspective to offer a deeper understanding of IRDA. We evaluate IRDA on five datasets, including a large-scale dataset, demonstrating its effectiveness in mitigating the adverse effects of imbalanced data distribution and its adaptability to various regression tasks.

Multi‐view synergistic enhanced fault recording data for transmission line fault classification

AbstractFault recorded data has been proven to be effective for fault diagnosis of overhead transmission lines. Utilizing deep learning to mine potential fault patterns in fault recording data is an inevitable trend. However, it is usually difficult to obtain massive labeled fault recording data, which results in deep learning‐based fault diagnosis models not being adequately trained. Although data augmentation methods provide ideas for expanding the training data, existing data augmentation algorithms (e.g. random perturbation‐based augmentation) may lead to distortion of multi‐view data, that is, time domain data and frequency domain data of the fault recorded data, which results in the inconsistency of physical properties and statistical distributions of the generated data and the actual recording data, and misguides the training of the models. Hence, this study proposes a transmission line fault classification method via the multi‐view synergistic enhancement of fault recording data. The methodology proposes to start with a synergistic enhancement of multi‐view data such as time and frequency domains of fault recording data, and utilizes contrastive learning to further improve the performance of the fault classification model while ensuring that the generated data is not distorted. Experimental results on three real‐world datasets validate the effectiveness of the proposed method.

Improved YOLOv8 for B-scan image flaw detection of the heavy-haul railway

With the high speed and heavy duty of railway transportation, internal flaw detection of railway rails has become a hot issue. Existing rail flaw detection systems have problems of low detection accuracy and occasional missed flaw detection. In this paper, a high-precision flaw detection based on data augmentation and YOLOv8 improvement is proposed. Firstly, three data augmentation algorithms based on the characteristics of B-scan images are designed to enrich the dataset of rail flaws. Then, the small target detection layer and the cross-layer connectivity module are added to capture more information for small targets. Finally, the introduction of dynamic weights to coordinate attention can adjust the attentional weights and capture long-range information. The experimental results show that the mAP50 of the model after data enhancement and algorithm improvement is 97.9%, which is improved by 4.4% from the baseline model, and the frame per second is 64.52. The proposed method effectively detects many typical flaws, including the railhead flaw, rail jaw flaw, screw hole crack, and bottom flaw, which can provide technology supports for on-site maintenance staff.

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.

Houseplant leaf classification system based on deep learning algorithms

AbstractBotanical experts are typically relied upon to classify houseplants since even subtle differences in characteristics such as leaves can distinguish one species from another. Therefore, an automated system for recognizing houseplant leaves with accuracy and reliability becomes a valuable asset for the identification of indoor plant species. In this paper, a houseplant leaf classification system utilizing deep learning algorithms is proposed, which has been improved to effectively classify and identify a variety of houseplant leaf types. The system uses the ResNet-50 architecture based on convolutional neural network to analyze features of the leaf images and extract relevant information for classification. In addition, this work presents a newly constructed local dataset consisting of 2500 images to classify species of houseplant leaves. The dataset includes ten types of houseplant leaves that are suitable for cultivation in various climates at home. The dataset was augmented using data augmentation algorithms to increase its size and reduce overfitting. The developed system was training and testing using a local dataset. To evaluate the improved model, comparative experiments were conducted utilizing pre-trained models (original ResNet-50 and MobileNet_v2). The improved model revealed recognition accuracy of 99% with the augmented dataset and 98.60% without the augmentation, affirming its effectiveness. The improved model could potentially be used in various fields, including horticulture, plant pathology, and environmental monitoring to identify plant species.

Effective time-series Data Augmentation with Analytic Wavelets for bearing fault diagnosis

In the realm of rotary machine maintenance, rolling bearings emerge as crucial yet frequently vulnerable components. Ensuring their operational integrity is pivotal for the overall safety and performance of the machinery. Traditional diagnostic methods have utilized a variety of signal processing techniques, ranging from blind source separation to wavelet analysis, and more recently, deep neural networks (DNNs). However, a significant challenge in this domain is the scarcity of labeled failure data in real-world applications, leading to a reliance on data augmentation (DA) techniques to enhance the quality and quantity of the training data. Compounding this problem is the use of DA techniques tailored for images, which are not suitable for the time–frequency representations. This study proposes an innovative Data Augmentation with Analytic Wavelets (DAAW) approach, tailored specifically for time-series (TS) data classification in bearing fault diagnosis. The essence of DAAW lies in its ability to generate synthetic scalogram samples that closely mirror the properties of original samples. The proposed DAAW is directly applied at the input stage of a learning model by successively adjusting predefined parameters to generate scalograms, without the need for auxiliary DA algorithms. This technique is versatile in generating synthetic time–frequency data samples, and the parameterization enables it to be tailored to different time series datasets. Through experimentation with publicly available datasets, we show that combining a simple Convolutional Neural Network (CNN) with the proposed approach results in improved performance. It outperforms existing methods, achieving a mean classification accuracy of 99.49%. The critical distance for average accuracy, as determined by the Nemenyi post-hoc test, is 0.4619. Our results also show enhanced performance with accuracy rates of 90.14%, 78.94%, and 62.86% even under low signal-to-noise ratio (SNR) conditions, using 80%, 60%, and 50% of the dataset, respectively, for model training. Notably, its efficacy remains robust even under challenging low SNR conditions. The proposed DAAW significantly improves classification accuracy in bearing fault diagnosis, simultaneously obviating the need for proxy DA tasks.

Convergence analysis of data augmentation algorithms for Bayesian robust multivariate linear regression with incomplete data

Gaussian mixtures are commonly used for modeling heavy-tailed error distributions in robust linear regression. Combining the likelihood of a multivariate robust linear regression model with a standard improper prior distribution yields an analytically intractable posterior distribution that can be sampled using a data augmentation algorithm. When the response matrix has missing entries, there are unique challenges to the application and analysis of the convergence properties of the algorithm. Conditions for geometric ergodicity are provided when the incomplete data have a “monotone” structure. In the absence of a monotone structure, an intermediate imputation step is necessary for implementing the algorithm. In this case, we provide sufficient conditions for the algorithm to be Harris ergodic. Finally, we show that, when there is a monotone structure and intermediate imputation is unnecessary, intermediate imputation slows the convergence of the underlying Monte Carlo Markov chain, while post hoc imputation does not. An R package for the data augmentation algorithm is provided.

Data Augmentation for Bayesian Deep Learning

Deep Learning (DL) methods have emerged as one of the most powerful tools for functional approximation and prediction. While the representation properties of DL have been well studied, uncertainty quantification remains challenging and largely unexplored. Data augmentation techniques are a natural approach to provide uncertainty quantification and to incorporate stochastic Monte Carlo search into stochastic gradient descent (SGD) methods. The purpose of our paper is to show that training DL architectures with data augmentation leads to efficiency gains. We use the theory of scale mixtures of normals to derive data augmentation strategies for deep learning. This allows variants of the expectation-maximization and MCMC algorithms to be brought to bear on these high dimensional nonlinear deep learning models. To demonstrate our methodology, we develop data augmentation algorithms for a variety of commonly used activation functions: logit, ReLU, leaky ReLU and SVM. Our methodology is compared to traditional stochastic gradient descent with back-propagation. Our optimization procedure leads to a version of iteratively re-weighted least squares and can be implemented at scale with accelerated linear algebra methods providing substantial improvement in speed. We illustrate our methodology on a number of standard datasets. Finally, we conclude with directions for future research.

Improving the prediction of extreme wind speed events with generative data augmentation techniques

Extreme Wind Speed events (EWS) are responsible for the worst damages caused by wind in wind farms. An accurate estimation of the frequency and intensity of EWS is essential to avoid wind turbine damage and to minimize cut-out events in these facilities. In this paper we discuss how generative Data Augmentation (DA) techniques improve the performance of Machine Learning (ML) and Deep Learning (DL) algorithms in EWS prediction problems. These problems are usually tackled as classification tasks, which are highly unbalanced due to the small number of EWS events in wind farms. Different versions of Variational AutoEncoders (VAE) are proposed and analysed in this work (VAEs, Conditional VAEs (CVAEs) and Class-Informed VAEs (CI-VAE)) as generative DA techniques to balance the data in EWS problems, leading to better performance of the prediction systems. The proposed generative DA techniques have been compared against traditional DA algorithms in a real problem of EWS prediction in Spain, considering ERA5 reanalysis data as predictive variables. The results showed that the CI-VAE with a Convolutional Neural Network approach obtained the best results, with values of Precision 0.62, Recall 0.74 and F1 score 0.67, improving up to 4% the results of the method without data augmentation techniques.

ALAD-YOLO:an lightweight and accurate detector for apple leaf diseases.

Suffering from various apple leaf diseases, timely preventive measures are necessary to take. Currently, manual disease discrimination has high workloads, while automated disease detection algorithms face the trade-off between detection accuracy and speed. Therefore, an accurate and lightweight model for apple leaf disease detection based on YOLO-V5s (ALAD-YOLO) is proposed in this paper. An apple leaf disease detection dataset is collected, containing 2,748 images of diseased apple leaves under a complex environment, such as from different shooting angles, during different spans of the day, and under different weather conditions. Moreover, various data augmentation algorithms are applied to improve the model generalization. The model size is compressed by introducing the Mobilenet-V3s basic block, which integrates the coordinate attention (CA) mechanism in the backbone network and replacing the ordinary convolution with group convolution in the Spatial Pyramid Pooling Cross Stage Partial Conv (SPPCSPC) module, depth-wise convolution, and Ghost module in the C3 module in the neck network, while maintaining a high detection accuracy. Experimental results show that ALAD-YOLO balances detection speed and accuracy well, achieving an accuracy of 90.2% (an improvement of 7.9% compared with yolov5s) on the test set and reducing the floating point of operations (FLOPs) to 6.1 G (a decrease of 9.7 G compared with yolov5s). In summary, this paper provides an accurate and efficient detection method for apple leaf disease detection and other related fields.

Comparison of the performance of different one-dimensional convolutional neural network models-based near-infrared spectra for determination of chlorpyrifos residues in corn oil

This study introduces a fast analytical approach for detecting chlorpyrifos residues in corn oil, based on a one-dimensional convolutional neural network (1D-CNN) structure for near-infrared (NIR) spectral deep learning models. Data augmentation algorithms were introduced to effectively expand the database of corn oil samples obtained in the experiment, ensuring sufficient training of the deep learning network. A well-designed 1D-CNN structure was implemented to create a multivariate calibration model. This model can be used to quantify chlorpyrifos residues in corn oil. The study found that using data augmentation with a 1D-CNN model led to improved performance compared to using the original spectra calibration in the model. Particularly, the 1D-CNN model’s best overall performance was achieved using data augmented with varying degrees of signal-to-noise ratio (SNR) algorithms, with a prediction coefficient of determination (RP2) of 0.9492 and a relative percent deviation (RPD) of 4.4569. The study demonstrates the feasibility of using NIR spectral deep learning models for quantitative analysis of pesticide residues in edible oils. Furthermore, data augmentation methods can effectively improve the accuracy of deep learning algorithms in NIR spectral chemometrics model calibration.

Anomaly-GAN: A data augmentation method for train surface anomaly detection

Train surface anomaly detection is an essential task in vision-based railway safety inspection. Although existing deep learning methods show great potential, their anomaly detection accuracy is affected by the lack of abnormal images. The number of abnormal images on the train surface is far less than that of normal images. One effective way to solve this problem is to expand the abnormal sample. However, most of train surface anomaly image generation methods faces difficulties in producing images with high quality and rich diversity at the same time. In addition, generating small-area anomalies is not ideal. An Anomaly-GAN based on mask pool, abnormal aware loss, and local versus global discriminators is thus proposed in this paper to solve these problems. The mask pool consists of prior-knowledge-based masks and expert-experience-based masks that guide model in generating anomalies with different shapes, rotation angles, spatial locations, and part numbers. The anomaly aware loss focuses on small-area anomalies, thereby promoting the generated anomalies with more detailed textures and richer semantics. The local versus global consistency discriminators combine local and global feature expressions that lead to the generation of more realistic and natural abnormal samples. The experiments show that, compared with other advanced data augmentation algorithms, the images generated by Anomaly-GAN achieve the best FID and LPIPS scores in all anomaly categories. In addition, compared with the case without data augmentaion, the proposed data enhancement method improves the performance of CNN on mAP and mIOU by 25.6% and 24.2%, respectively. Test code is available in https://github.com/AI-Dream-Chaser/Anomaly-GAN.

Enhancing Pavement Distress Detection Using a Morphological Constraints-Based Data Augmentation Method

Pavement distress data in a single section usually presents a long-tailed distribution, with potholes, sealed cracks, and other distresses normally located at the tail. This distribution will seriously affect the performance and robustness of big data-driven deep learning detection models. Conventional data augmentation algorithms only expand the amount of data by image transformation and fail to enlarge the data diversity. Due to such a drawback, this paper proposes a novel two-stage pavement distress image augmentation pattern, in which a mask is generated randomly according to the geometric features of the distress in the first stage; and in the second stage, a distress-free pavement image with the fused mask is transformed into a pavement distress image. Furthermore, two convolutional networks, M-DCGAN and MDTMN, are designed to complete the generation task in two stages separately. In comparison with other generation algorithms, the quality and diversity of the generation results of proposed algorithms are better than other algorithms. In addition, distress detection tests are conducted which indicate that the expanded dataset can raise the IoU from 48.83% to 83.65% at maximum, and the augmented data by the proposed algorithm contributes more to the detection performance.

Deep Learning and Transfer Learning Methods to Effectively Diagnose Cervical Cancer from Liquid-Based Cytology Pap Smear Images

As cervical cancer is considered one of the leading causes of death for women globally, different screening techniques have emerged. As the Papanicolaou technique generates high numbers of false negatives due to only testing 20% of a sample, the liquid-based cytology technique was developed to test 100% of the sample and improve accuracy. However, as the larger sample size has made it difficult to detect the lesion images through a microscope, studies have looked for ways to intelligently analyze sample. The aim of this study is to develop an artificial intelligence image recognition system that detects the lesion level of cervical cancer of liquid-based Pap smears under the Bethesda classification of cancer (NI/LSIEL/HSIEL/SCC). For this purpose, six activities were carried out: dataset selection, data augmentation, optimization, model development, evaluation and system construction. A dataset built from publicly available Pap smear images and passed through data augmentation algorithms generated a total of 2,676 images. Two models, ResNet50V2 and ResNet101V2, were developed under Deep Learning and Transfer Learning protocols. The evaluation showed that the ResNet50V2 model obtained better performance, where the classification of HSIL and SCC type images obtained a precision of 0.98 and achieved an accuracy of 0.97. Finally, the system based on the ResNet50V2 model was built and its performance was validated.

ПРИМЕНЕНИЕ ГЛУБОКОГО ОБУЧЕНИЯ ДЛЯ АУГМЕНТАЦИИ И ГЕНЕРАЦИИ ПОДВОДНОГО НАБОРА ДАННЫХ С ПРОМЫШЛЕННЫМИ ОБЪЕКТАМИ

The purpose of the study: development of a deep learning method for augmentation and generation of a problem-oriented dataset containing industrial objects, including the development of more efficient data generation and augmentation algorithms based on deep learning, which allow you to create more diverse and realistic data corresponding to industrial objects that can be transferred from one style domain to another. The goal set in the study is related to the actual scientific and technical problem of providing computer vision in systems operating in the underwater environment. These can be autonomous uninhabited underwater vehicles that look for breaks in pipelines, analyze oil leaks, the movement of schools of fish, etc. However, today there is not enough data containing the described objects in the conditions of their real existence. Thus, it is necessary to provide the training sample with realistic images. Research methods: the CycleGAN architecture, which converts a dataset containing images of various objects taken in a laboratory or in a conventional aboveground environment into a dataset containing the cha¬racteristics of an underwater environment. To evaluate the developed augmentation algorithm, it is proposed to use image classification by domains, which can be performed using the ResNet convolutional neural network. Results of the study. A tool is presented to solve the problem of the lack of underwater datasets, a deep learning model is developed, which is used to create images with underwater elements. The model works on the principle of a cyclic generative adversarial network, which receives a real image of an industrial facility in surface conditions as an input, and returns a generated image of the same industrial facility in underwater conditions as an output. Visual analysis of images shows that this method is quite adequate. In addition, a test on the classification model showed almost 100% ability of the neural network to distinguish between domains. Conclusion. The study showed that the CycleGAN model can be used to create images of various objects in the underwater environment. In the future, it is possible to search for additional augmentation procedures, in addition, augmentations of the generated set can be used images, which will also provide researchers and developers with sufficient material with industrial facilities in the underwater environment. This can improve the quality of developments.

Time Series of Counts under Censoring: A Bayesian Approach.

Censored data are frequently found in diverse fields including environmental monitoring, medicine, economics and social sciences. Censoring occurs when observations are available only for a restricted range, e.g., due to a detection limit. Ignoring censoring produces biased estimates and unreliable statistical inference. The aim of this work is to contribute to the modelling of time series of counts under censoring using convolution closed infinitely divisible (CCID) models. The emphasis is on estimation and inference problems, using Bayesian approaches with Approximate Bayesian Computation (ABC) and Gibbs sampler with Data Augmentation (GDA) algorithms.

Improving Many-to-Many Neural Machine Translation via Selective and Aligned Online Data Augmentation

Multilingual neural machine translation (MNMT) models are theoretically attractive for low- and zero-resource language pairs with the impact of cross-lingual knowledge transfer. Existing approaches mainly focus on English-centric directions and always underperform compared to their pivot-based counterparts for non-English directions. In this work, we aim to build a many-to-many MNMT system with an emphasis on the quality of non-English directions by exploring selective and aligned online data augmentation algorithms. Based on our findings showing that the augmented synthetic samples are not “the more, the better” we propose selective online back-translation (SOBT) and thoroughly study different selection criteria to pick suitable samples for training. Furthermore, we boost SOBT with cross-lingual online substitution (CLOS) to align token representations and encourage transfer learning. Our intuition is based on the hypothesis that a universal cross-lingual representation leads to a better multilingual translation performance, especially for non-English directions. Comparing to previous state-of-the-art many-to-many MNMT models and conventional pivot-based methods, experiments on IWSLT2014 and OPUS-100 translation benchmarks show that our approach achieves a competitive or even better performance on English-centric directions and achieves up to ∼12 BLEU for non-English directions. All of our models and codes are publicly available.

A survey of automated data augmentation algorithms for deep learning-based image classification tasks

In recent years, one of the most popular techniques in the computer vision community has been the deep learning technique. As a data-driven technique, deep model requires enormous amounts of accurately labelled training data, which is often inaccessible in many real-world applications. A data-space solution is Data Augmentation (DA), that can artificially generate new images out of original samples. Image augmentation strategies can vary by dataset, as different data types might require different augmentations to facilitate model training. However, the design of DA policies has been largely decided by the human experts with domain knowledge, which is considered to be highly subjective and error-prone. To mitigate such problem, a novel direction is to automatically learn the image augmentation policies from the given dataset using Automated Data Augmentation (AutoDA) techniques. The goal of AutoDA models is to find the optimal DA policies that can maximize the model performance gains. This survey discusses the underlying reasons of the emergence of AutoDA technology from the perspective of image classification. We identify three key components of a standard AutoDA model: a search space, a search algorithm and an evaluation function. Based on their architecture, we provide a systematic taxonomy of existing image AutoDA approaches. This paper presents the major works in AutoDA field, discussing their pros and cons, and proposing several potential directions for future improvements.

The Status and Challenges of Image Data Augmentation Algorithms

Image data-augmentation algorithms effectively trump the problem of insufficient training samples for deep learning in some application fields, and it is typically for scholars to choose some of them for various computer vision tasks. But as the algorithms develop rapidly, the early proposed classification that the data-augmentation algorithms are sorted into classical ways and generating methods is no more suitable, because such classification misses some other meaningful strategies. Besides, it is frustrating for someone to decide which is the exact method to undertake, though there are too many optional algorithms to choose. Towards the goal of making some suggestions, the paper categorizes image data-augmentation algorithms into three kinds from the perspective of algorithm strategy, and they are matrix transformation algorithm, feature expansion algorithm and model generation based on neural network algorithm. The paper analyzes the typical algorithm principle, performance, application scenarios, research status and future challenges, and forecasts the development trend of data augmentation algorithms. The paper can provide academic reference for data augmentation algorithm in the fields of medicine and military.

An adversarial domain adaptation approach combining dual domain pairing strategy for IoT intrusion detection under few-shot samples

The security of the Internet of Things (IoT) is an integral research topic to support its applicability. With the rapid development of IoT technology, security threats have become increasingly serious. Attackers turn their attention to large IoT devices with many security vulnerabilities, and many infected IoT devices are used to form botnets to launch intrusions. Furthermore, IoT communication networks poses security risks. Therefore, it is necessary to design a model for the IoT intrusion detection. However, the current models are trained using big data samples. Moreover, IoT devices can only intercept a few-shot samples in special scenarios, resulting in lower efficiency and accuracy. In this study, we propose an adversarial domain adaptation (ADDA) approach with dual domain pairing strategy for IoT intrusion detection under few-shot samples. First, we design two data augmentation algorithms to construct a sample transfer model that conforms to the joint dual domain distribution. We then combine the dual domain pairing strategy and ADDA model to perform supervised domain adaptation. In addition, a lightweight network is designed to extract features. Finally, a new loss function and domain discriminator model are constructed to achieve IoT intrusion detection under few-shot samples. Different results demonstrate the effectiveness of our model.