Effect Of Data Augmentation Research Articles

An improved deep convolutional generative adversarial network for quantification of catechins in fermented black tea

The rapid and non-destructive quantification of catechins in fermented black tea is crucial for evaluating the quality of black tea. The combination of hyperspectral imaging and chemometrics has been applied for quantitative detection, but its performance is usually constrained by the limited dataset size. Targeted at the challenge of insufficient samples in regression analysis of catechins, this study proposes an improved deep convolutional generative adversarial network with labeling module, named as DCGAN-L for hyperspectral data augmentation. The DCGAN-L consists of the spectral and label generating modules. First the synthetic spectra were generated, and an indicator was proposed to evaluate their quality. Then, the corresponding label values were generated, including epicatechin gallate (ECG), epicatechin (EGC), catechin (C), and total catechin (CC). For label generating, the Euclidean distances between the synthetic spectrum and all true spectra were measured, followed by allocating weights for calculating the label values based on these distances. Subsequently, the training dataset was augmented with the generated synthetic data. The effect of data augmentation was finally evaluated based on two regression models of random forest (RF) and broad learning system (BLS) for the quantification of catechins. Compared with the results before data augmentation, the average R2 of RF and BLS models increased by 0.044 and 0.164, respectively. The proposed DCGAN-L model allows for the rapid, non-destructive quantification of catechins in black tea in the case of limited sample size.

Effectiveness of data-augmentation on deep learning in evaluating rapid on-site cytopathology at endoscopic ultrasound-guided fine needle aspiration

Rapid on-site cytopathology evaluation (ROSE) has been considered an effective method to increase the diagnostic ability of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA); however, ROSE is unavailable in most institutes worldwide due to the shortage of cytopathologists. To overcome this situation, we created an artificial intelligence (AI)-based system (the ROSE-AI system), which was trained with the augmented data to evaluate the slide images acquired by EUS-FNA. This study aimed to clarify the effects of such data-augmentation on establishing an effective ROSE-AI system by comparing the efficacy of various data-augmentation techniques. The ROSE-AI system was trained with increased data obtained by the various data-augmentation techniques, including geometric transformation, color space transformation, and kernel filtering. By performing five-fold cross-validation, we compared the efficacy of each data-augmentation technique on the increasing diagnostic abilities of the ROSE-AI system. We collected 4059 divided EUS-FNA slide images from 36 patients with pancreatic cancer and nine patients with non-pancreatic cancer. The diagnostic ability of the ROSE-AI system without data augmentation had a sensitivity, specificity, and accuracy of 87.5%, 79.7%, and 83.7%, respectively. While, some data-augmentation techniques decreased diagnostic ability, the ROSE-AI system trained only with the augmented data using the geometric transformation technique had the highest diagnostic accuracy (88.2%). We successfully developed a prototype ROSE-AI system with high diagnostic ability. Each data-augmentation technique may have various compatibilities with AI-mediated diagnostics, and the geometric transformation was the most effective for the ROSE-AI system.

An improved data augmentation approach and its application in medical named entity recognition

Performing data augmentation in medical named entity recognition (NER) is crucial due to the unique challenges posed by this field. Medical data is characterized by high acquisition costs, specialized terminology, imbalanced distributions, and limited training resources. These factors make achieving high performance in medical NER particularly difficult. Data augmentation methods help to mitigate these issues by generating additional training samples, thus balancing data distribution, enriching the training dataset, and improving model generalization. This paper proposes two data augmentation methods—Contextual Random Replacement based on Word2Vec Augmentation (CRR) and Targeted Entity Random Replacement Augmentation (TER)—aimed at addressing the scarcity and imbalance of data in the medical domain. When combined with a deep learning-based Chinese NER model, these methods can significantly enhance performance and recognition accuracy under limited resources. Experimental results demonstrate that both augmentation methods effectively improve the recognition capability of medical named entities. Specifically, the BERT-BiLSTM-CRF model achieved the highest F1 score of 83.587%, representing a 1.49% increase over the baseline model. This validates the importance and effectiveness of data augmentation in medical NER.

Effect of Data Augmentation, Cross-Validation Methods in Robustness of Explainable Speech Based Emotion Recognition

Effect of Data Augmentation, Cross-Validation Methods in Robustness of Explainable Speech Based Emotion Recognition

Effect of Data Augmentation Using Deep Learning on Predictive Models for Geopolymer Compressive Strength

In recent years, machine learning models have become a potential approach in accurately predicting the concrete compressive strength, which is essential for the real-world application of geopolymer concrete. However, the precursor system of geopolymer concrete is known to be more heterogeneous compared to Ordinary Portland Cement (OPC) concrete, adversely affecting the data generated and the performance of the models. To its advantage, data enrichment through deep learning can effectively enhance the performance of prediction models. Therefore, this study investigates the capability of tabular generative adversarial networks (TGANs) to generate data on mixtures and compressive strength of geopolymer concrete. It assesses the impact of using synthetic data with various models, including tree-based, support vector machines, and neural networks. For this purpose, 930 instances with 11 variables were collected from the open literature. In particular, 10 variables including content of fly ash, slag, sodium silicate, sodium hydroxide, superplasticizer, fine aggregate, coarse aggregate, added water, curing temperature, and specimen age are considered as inputs, while compressive strength is the output of the models. A TGAN was employed to generate an additional 1000 data points based on the original dataset for training new predictive models. These models were evaluated on real data test sets and compared with models trained on the original data. The results indicate that the developed models significantly improve performance, particularly neural networks, followed by tree-based models and support vector machines. Moreover, data characteristics greatly influence model performance, both before and after data augmentation.

Quantitative and Visual Analysis of Data Augmentation and Hyperparameter Optimization in Deep Learning-Based Segmentation of Low-Grade Glioma Tumors Using Grad-CAM.

This study executes a quantitative and visual investigation on the effectiveness of data augmentation and hyperparameter optimization on the accuracy of deep learning-based segmentation of LGG tumors. The study employed the MobileNetV2 and ResNet backbones with atrous convolution in DeepLabV3+ structure. The Grad-CAM tool was also used to interpret the effect of augmentation and network optimization on segmentation performance. A wide investigation was performed to optimize the network hyperparameters. In addition, the study examined 35 different models to evaluate different data augmentation techniques. The results of the study indicated that incorporating data augmentation techniques and optimization can improve the performance of segmenting brain LGG tumors up to 10%. Our extensive investigation of the data augmentation techniques indicated that enlargement of data from 90° and 225° rotated data,up to down and left to right flipping are the most effective techniques. MobilenetV2 as the backbone,"Focal Loss" as the loss function and "Adam" as the optimizer showed the superior results. The optimal model (DLG-Net) achieved an overall accuracy of 96.1% with a loss value of 0.006. Specifically, the segmentation performance for Whole Tumor (WT), Tumor Core (TC), and Enhanced Tumor (ET) reached a Dice Similarity Coefficient (DSC) of 89.4%, 70.1%, and 49.9%, respectively. Simultaneous visual and quantitative assessment of data augmentation and network optimization can lead to an optimal model with a reasonable performance in segmenting the LGG tumors.

Layer Configurations of BERT for Multitask Learning and Data Augmentation

Multitask learning (MTL) and data augmentation are becoming increasingly popular in natural language processing (NLP). These techniques are particularly useful when data are scarce. In MTL, knowledge learned from one task is applied to another. To address data scarcity, data augmentation facilitates by providing additional synthetic data during model training. In NLP, the bidirectional encoder representations from transformers (BERT) model is the default candidate for various tasks. MTL and data augmentation using BERT have yielded promising results. However, a detailed study regarding the effect of using MTL in different layers of BERT and the benefit of data augmentation in these configurations has not been conducted. In this study, we investigate the use of MTL and data augmentation from generative models, specifically for category classification, sentiment classification, and aspect-opinion sequence-labeling using BERT. The layers of BERT are categorized into top, middle, and bottom layers, which are frozen, shared, or unshared. Experiments are conducted to identify the optimal layer configuration for improved performance compared with that of single-task learning. Generative models are used to generate augmented data, and experiments are performed to reveal their effectiveness. The results indicate the effectiveness of the MTL configuration compared with single-task learning as well as the effectiveness of data augmentation using generative models for classification tasks.

Effect of Data Augmentation on Deep-Learning-Based Segmentation of Long-Axis Cine-MRI

Automated Cardiac Magnetic Resonance segmentation serves as a crucial tool for the evaluation of cardiac function, facilitating faster clinical assessments that prove advantageous for both practitioners and patients alike. Recent studies have predominantly concentrated on delineating structures on short-axis orientation, placing less emphasis on long-axis representations due to the intricate nature of structures in the latter. Taking these consideration into account, we present a robust hierarchy-based augmentation strategy coupled with the compact and fast Efficient-Net (ENet) architecture for the automated segmentation of two-chamber and four-chamber Cine-MRI images. We observed an average Dice improvement of 0.99% on the two-chamber images and of 2.15% on the four-chamber images, and an average Hausdorff distance improvement of 21.3% on the two-chamber images and of 29.6% on the four-chamber images. The practical viability of our approach was validated by computing clinical metrics such as the Left Ventricular Ejection Fraction (LVEF) and left ventricular volume (LVC). We observed acceptable biases, with a +2.81% deviation on the LVEF for the two-chamber images and a +0.11% deviation for the four-chamber images.

Effects of data augmentation by replicating instances : Classification performance by Decision trees

Numerous medical data have classification problems for it has unbalanced instances. This means the medical data has less number of positive instances because it refers to disease diagnosis. Even the world's common disease diabetes data is unbalanced. We have the machine learning algorithms to combat the imbalanced classification. Among them, resampling is a useful technique that helps to balance the instance based on the classes’ majority and minority by under-sampling and oversampling methods. However, in spite of its circulation, the sampling has issues in the efficient evaluation of small-sized data. This study measured sampling methods and machine learning classifiers of different split percentages using a diabetes dataset which waver units of imbalance and produce better accuracy. The evaluation measure for each replication percentage is calculated and the same is interpreted in the Discussion.

In-Domain Transfer Learning Strategy for Tumor Detection on Brain MRI.

Transfer learning has gained importance in areas where there is a labeled data shortage. However, it is still controversial as to what extent natural image datasets as pre-training sources contribute scientifically to success in different fields, such as medical imaging. In this study, the effect of transfer learning for medical object detection was quantitatively compared using natural and medical image datasets. Within the scope of this study, transfer learning strategies based on five different weight initialization methods were discussed. A natural image dataset MS COCO and brain tumor dataset BraTS 2020 were used as the transfer learning source, and Gazi Brains 2020 was used for the target. Mask R-CNN was adopted as a deep learning architecture for its capability to effectively handle both object detection and segmentation tasks. The experimental results show that transfer learning from the medical image dataset was found to be 10% more successful and showed 24% better convergence performance than the MS COCO pre-trained model, although it contains fewer data. While the effect of data augmentation on the natural image pre-trained model was 5%, the same domain pre-trained model was measured as 2%. According to the most widely used object detection metric, transfer learning strategies using MS COCO weights and random weights showed the same object detection performance as data augmentation. The performance of the most effective strategies identified in the Mask R-CNN model was also tested with YOLOv8. Results showed that even if the amount of data is less than the natural dataset, in-domain transfer learning is more efficient than cross-domain transfer learning. Moreover, this study demonstrates the first use of the Gazi Brains 2020 dataset, which was generated to address the lack of labeled and qualified brain MRI data in the medical field for in-domain transfer learning. Thus, knowledge transfer was carried out from the deep neural network, which was trained with brain tumor data and tested on a different brain tumor dataset.

CNN-based in situ tool wear detection: A study on model training and data augmentation in turning inserts

We present an automated turning insert wear detection system developed for aeronautical Low Pressure Turbines (LPT) casing machining process based on a binary classifier using Convolutional Neural Networks (CNN). This method involves acquiring the image on the machine itself. During this process, removing the insert from the tool holder is not necessary, and the wear assessment is performed before the next workpiece is mechanized. Since datasets in tool wear prediction are often imbalanced, a multi perspective camera technology as well as data augmentation and class weighting are utilized to address both the number of worn parts considered and the cost of image acquisition. In this study four different insert types and two CNN architectures (specific and universal models) are considered and evaluated. The effects of data augmentation and training set size are discussed. While the models trained perform well on round inserts, they fail on rhombic insert types. An accuracy up to 97.8% (Matthew’s correlation coefficient of 0.955) is achieved by the machine learning model. Additionally, it can detect defects on a variety of insert types.

Multi-class classification of brain tumor types from MR images using EfficientNets

Accurate classification of the type of brain tumor plays an important role in the early diagnosis of the tumor which can be the difference between life and death. Magnetic Resonance Imaging (MRI) is commonly used to capture high-contrast grayscale images of the brain and is a non-invasive method for brain tumor diagnosis. Deep Learning (DL) using Convolutional Neural Networks (CNN) has revolutionized Computer Aided Diagnostic (CAD) systems by producing remarkable results for various medical imaging analysis tasks including brain tumor detection. In this research, a transfer learning-based fine-tuning approach for the classification of brain tumors into three classes i.e. glioma, meningioma, and pituitary tumor using EfficientNets is carried out. Five variants of pre-trained models from the EfficientNets family i.e. EfficientNetB0 – EfficientNetB4 are fine-tuned on publically available CE-MRI “Figshare – Brain Tumor Dataset”. The proposed method of fine-tuning pre-trained EfficientNet is carried out by first loading ImageNet weights to the EfficientNet model followed by the addition of several top layers and a fully connected layer for the classification of brain tumor types. Several experiments are carried out to analyze the robustness of the proposed fine-tuned EfficientNets in comparison to other pre-trained models. Moreover, the effect of data augmentation on the model’s test accuracy is also studied. Finally, the Grad-CAM visualization of the attention maps of the best model is presented that successfully highlights the tumorous region of the brain cell. The proposed method of fine-tuning pre-trained EfficientNet showed significant performance using EfficientNetB2 as a backbone achieving overall test accuracy, precision, recall/sensitivity, and F1-score of 98.86%, 98.65%, 98.77%, and 98.71% respectively. The proposed fine-tuned EfficientNetB2 is lightweight, computationally inexpensive (during training and inference), and generalizes well. The source code is available at https://github.com/FatimaZulfiqar/multi-class-brain-tumor-classification.

An augmented customized deep learning approach for brain tumour identification

ABSTRACT Brain is one of the crucial organs in the human body and the survival rate for those who are affected by brain tumours across the globe is very low. There might be low survival rate yet it can be improved by identifying the disease by using MRI of the brain at a very early stage. At this juncture, the automatic identification of the tumour accurately using MRI images is very essential. Once the deep learning came into existence, the accuracy of identification of various biomedical diseases using MRI or X-ray images has been improved.Existing Systems are using deep learning approach identification of the diseases but they are lacking in the size of the data which they are using for the implementation purpose. Considering this aspect as an inspiration, a framework proposed for the identification of brain tumour with a customized neural network along with a CNN architecture. The dataset considered for the implementation of this framework is an open dataset obtained from Kaggle. As the dataset is smaller, the data augmentation technique is used to improve the dataset size. So, the effect of data augmentation in attaining the accuracy is also discussed. The training accuracy obtained during pre-data augmentation is about 85.76% and post-data augmentation is about 97.85%.

Exploring question generation in medical intelligent system using entailment

The concept of a medical intelligent system has steadily garnered attention as modern technology advances. An intelligent medical system is a medical system that develops a certain amount of intelligence and performs a task without the assistance of a human. A chatbot can be thought of as a medical intelligent consultation system. The chatbot's question generation quality can be improved by creating more relevant questions depending on the patient's demands. Question generation, in addition to chatbots, is used to assess a learner's comprehension. This paper proposes a two-step approach to question generation. The first stage generates the entailment for the sentence that the question should be generated for. The generated entailed sentences are used to create questions in the second step. By generating questions from the original sentence, one can discover relevant information about the sentence. Furthermore, to increase the size of the entailment dataset, a data augmentation approach is used in this paper. The proposed work in this paper focuses on the importance of entailment in question generation and also studies the influence of entailment on the questions generated. Since data augmentation is employed, the overall effectiveness of data augmentation on the model is also investigated.

Effects of Data Augmentation with the BNNSMOTE Algorithm in Seizure Detection Using 1D-MobileNet.

Automatic seizure detection technology has important implications for reducing the workload of neurologists for epilepsy diagnosis and treatment. Due to the unpredictable nature of seizures, the imbalanced classification of seizure and nonseizure data continues to be challenging. In this work, we first propose a novel algorithm named the borderline nearest neighbor synthetic minority oversampling technique (BNNSMOTE) to address the imbalanced classification problem and improve seizure detection performance. The algorithm uses the nearest neighbor notion to generate nonseizure samples near the boundary, then determines the seizure samples that are difficult to learn at the boundary, and lastly selects seizure samples at random to be used in the synthesis of new samples. In view of the characteristic that electroencephalogram (EEG) signals are one-dimensional signals, we then develop a 1D-MobileNet model to validate the algorithm's performance. Results demonstrate that the proposed algorithm outperforms previous seizure detection methods on the CHB-MIT dataset, achieving an average accuracy of 99.40%, a recall value of 87.46%, a precision of 97.17%, and an F1-score of 91.90%, respectively. We also had considerable success when we used additional datasets for verification at the same time. Our algorithm's data augmentation effects are more pronounced and perform better at seizure detection than the existing imbalanced techniques. Besides, the model's parameters and calculation volume have been significantly reduced, making it more suitable for mobile terminals and embedded devices.

Data Augmentation in 3D Object Detection for self-driving vehicles: the role of original and augmented training samples

Safe self-driving vehicles require precise 3D Object Identification. LiDAR sensors are key in accomplishing such a task, as LiDARs produce high-definition point clouds. Such point clouds are then processed by 3D Object Detection models to finally detect objects.Most Object Detection models require massive amounts of data to be trained. Gathering and processing this data is an expensive and time-consuming task, which is why the information taken from each sample must be fully harnessed. Such can be done through Data Augmentation. Data Augmentation contributes significantly for improving performance, being at least as relevant as the advances in the Object Detection models themselves.A few studies have been reported regarding the effectiveness of Data Augmentation. However, the role played by original and augmented samples has been neglected. This work reports the first-ever detailed quantification of the impact that the inclusion of original and augmented samples in a dataset has in 3D Object Detection in the context of autonomous driving. The obtained results show that although a good augmentation strategy is crucial to the model’s performance, it is only as good as the quality of the original samples allows it to be.

Attention Based Convolutional Neural Network with Multi-frequency Resolution Feature for Environment Sound Classification.

The environmental sound classification has great research significance in the fields of intelligent audio monitoring and other fields. A novel multi-frequency resolution (MFR) feature is proposed in this paper to solve the problem that the existing single frequency resolution time–frequency features of sound cannot effectively express the characteristics of multiple types of sound. The MFR feature is composed of three features with different frequency resolutions, which are compressed in varying degrees at the time dimension. This method not only has the effect of data augmentation but also can obtain more context information during the feature extraction. And the MFR features of Log-Mel Spectrogram, Cochleagram, and Constant Q-Transform are combined to form a multi-channel MFR feature. Also, a network named SacNet is built, which can effectively solve the problem that the time–frequency feature map of sound contains more invalid information. The basic structural unit of the SacNet consists of two parallel branches, one using depthwise separable convolution as the main feature extractor, and the other using spatial attention module to extract more effective information. Experiment results have demonstrated that the proposed method achieves the state-of-the-art accuracy of 97.5%, 93.1%, and 95.3% on three benchmark datasets of ESC10, ESC50, and UrbanSound8K respectively, which are increased by 3.3%, 0.5%, and 2.3% respectively compared with the previous advanced methods.

Effects of data augmentation and filters for automatic diagnosis of 12-lead ECG using residual network based deep learning

Effects of data augmentation and filters for automatic diagnosis of 12-lead ECG using residual network based deep learning

Deep learning for component fault detection in electricity transmission lines

Component fault detection and inventory are one of the most significant bottlenecks facing the electricity transmission and distribution utility establishments especially in developing countries for delivery of efficient services to the customers and to ensure proper asset audit and management for network optimization and load forecasting. For lack of technology and data, insecurity, the complexity associated with traditional methods, untimeliness, and general human cost, electricity assets monitoring, and management have remained a big problem in many developing countries. In view of this, we explored the use of oblique UAV imagery with high spatial resolution and fine-tuned deep Convolutional Neural Networks (CNNs) for automatic faulty component inspection and inventory in an Electric power transmission network (EPTN). This study investigated the capability of the Single Shot Multibox Detector (SSD), a one-stage object detection model on the electric transmission power line imagery to localize, detect and classify faults. Our proposed neural network model is a CNN based on a multiscale layer feature pyramid network (FPN) using aerial image patches and ground truth to localise and detect faults through a one-phase procedure. The SSD Rest50 architecture variation performed the best with a mean Average Precision (mAP) of 89.61%. All the developed SSD-based models achieve a high precision rate and low recall rate in detecting faulty components, thus achieving acceptable balance levels of F1-score and representation. We have established in this paper that combined use of UAV imagery and computer vision presents a low-cost method for easy and timely electricity asset inventory, especially in developing countries. This study also provides the guide to various considerations when adopting this technology in terms of the choice of deep learning architecture, adequate training samples over multiple fault characteristics, effects of data augmentation, and balancing of intra-class heterogeneity.

The Effect of Data Augmentation in Deep Learning Approach for Peripheral Blood Leukocyte Recognition.

Data augmentation is reported as a useful technique to generate a large amount of image datasets from a small image dataset. The aim of this study is to clarify the effect of data augmentation for leukocyte recognition with deep learning. We performed three different data augmentation methods (rotation, scaling, and distortion) as pretreatment on the original images. The subjects of clinical assessment were 51 healthy persons. The thin-layer blood smears were prepared from peripheral blood and stained with MG. The effect of data augmentation with rotation was the only significant effective technique in AI model generation for leukocyte recognition. On contrast, the effect of data augmentation with image distortion or image scaling was poor, and accuracy improvement was limited to specific leukocyte categories. Although data augmentation is one effective method for high accuracy in AI training, we consider that a highly effective method should be selected.