Data Augmentation Techniques Research Articles

A Novel Data Augmentation Method for Radiomics Analysis Using Image Perturbations.

Radiomics extracts hundreds of features from medical images to quantitively characterize a region of interest (ROI). When applying radiomics, imbalanced or small dataset issues are commonly addressed using under or over-sampling, the latter being applied directly to the extracted features. Aim of this study is to propose a novel balancing and data augmentation technique by applying perturbations (erosion, dilation, contour randomization) to the ROI in cardiac computed tomography images. From the perturbed ROIs, radiomic features are extracted, thus creating additional samples. This approach was tested addressing the clinical problem of distinguishing cardiac amyloidosis (CA) from aortic stenosis (AS) and hypertrophic cardiomyopathy (HCM). Twenty-one CA, thirty-two AS and twenty-one HCM patients were included in the study. From each original and perturbed ROI, 107 radiomic features were extracted. The CA-AS dataset was balanced using the perturbation-based method along with random over-sampling, adaptive synthetic (ADASYN) and the synthetic minority oversampling technique (SMOTE). The same methods were tested to perform data augmentation dealing with CA and HCM. Features were submitted to robustness, redundancy, and relevance analysis testing five feature selection methods (p-value, least absolute shrinkage and selection operator (LASSO), semi-supervised LASSO, principal component analysis (PCA), semi-supervised PCA). Support vector machine performed the classification tasks, and its performance were evaluated by means of a 10-fold cross-validation. The perturbation-based approach provided the best performances in terms of f1 score and balanced accuracy in both CA-AS (f1 score: 80%, AUC: 0.91) and CA-HCM (f1 score: 86%, AUC: 0.92) classifications. These results suggest that ROI perturbations represent a powerful approach to address both data balancing and augmentation issues.

Toward robust and high-throughput detection of seed defects in X-ray images via deep learning

BackgroundThe detection of internal defects in seeds via non-destructive imaging techniques is a topic of high interest to optimize the quality of seed lots. In this context, X-ray imaging is especially suited. Recent studies have shown the feasibility of defect detection via deep learning models in 3D tomography images. We demonstrate the possibility of performing such deep learning-based analysis on 2D X-ray radiography for a faster yet robust method via the X-Robustifier pipeline proposed in this article.Results2D X-ray images of both defective and defect-free seeds were acquired. A deep learning model based on state-of-the-art object detection neural networks is proposed. Specific data augmentation techniques are introduced to compensate for the low ratio of defects and increase the robustness to variation of the physical parameters of the X-ray imaging systems. The seed defects were accurately detected (F1-score >90%), surpassing human performance in computation time and error rates. The robustness of these models against the principal distortions commonly found in actual agro-industrial conditions is demonstrated, in particular, the robustness to physical noise, dimensionality reduction and the presence of seed coating.ConclusionThis work provides a full pipeline to automatically detect common defects in seeds via 2D X-ray imaging. The method is illustrated on sugar beet and faba bean and could be efficiently extended to other species via the proposed generic X-ray data processing approach (X-Robustifier). Beyond a simple proof of feasibility, this constitutes important results toward the effective use in the routine of deep learning-based automatic detection of seed defects.

Extensive data engineering to the rescue: building a multi-species katydid detector from unbalanced, atypical training datasets.

Passive acoustic monitoring (PAM) is a powerful tool for studying ecosystems. However, its effective application in tropical environments, particularly for insects, poses distinct challenges. Neotropical katydids produce complex species-specific calls, spanning mere milliseconds to seconds and spread across broad audible and ultrasonic frequencies. However, subtle differences in inter-pulse intervals or central frequencies are often the only discriminatory traits. These extremities, coupled with low source levels and susceptibility to masking by ambient noise, challenge species identification in PAM recordings. This study aimed to develop a deep learning-based solution to automate the recognition of 31 katydid species of interest in a biodiverse Panamanian forest with over 80 katydid species. Besides the innate challenges, our efforts were also encumbered by a limited and imbalanced initial training dataset comprising domain-mismatched recordings. To overcome these, we applied rigorous data engineering, improving input variance through controlled playback re-recordings and by employing physics-based data augmentation techniques, and tuning signal-processing, model and training parameters to produce a custom well-fit solution. Methods developed here are incorporated into Koogu, an open-source Python-based toolbox for developing deep learning-based bioacoustic analysis solutions. The parametric implementations offer a valuable resource, enhancing the capabilities of PAM for studying insects in tropical ecosystems. This article is part of the theme issue 'Towards a toolkit for global insect biodiversity monitoring'.

Chest Xray Medical Diagnosis with Deep Learning

Detecting Disease in Chest X-Rays with Deep Learning Chest X-rays (CXRs) is a common and vital diagnostic tool, but interpreting them accurately requires trained radiologists who can be scarce in many areas. This research explores the potential of deep learning (DL), a type of artificial intelligence (AI), to automatically detect diseases in CXRs. The review explains the basics of DL for CXR analysis, including how deep neural networks (DNNs) work and how transfer learning and data augmentation techniques can improve their performance. It then examines recent studies on applying DNNs to identify common CXR abnormalities like lung nodules, pneumonia, and pneumothorax. This includes exploring multi-class classification, where the model can detect multiple diseases simultaneously. The abstract compares the performance of these techniques with human observers and discusses the challenges of implementing DNN models in clinical practice, particularly their relationship with radiologists. Overall, the research investigates how DL can potentially improve disease detection using CXRs, addressing the need for accurate and accessible diagnostic tools. Key Words: Chest X-rays, Deep Learning, Disease Detection, Artificial Intelligence, Neural Networks

Assessing the reliability of point mutation as data augmentation for deep learning with genomic data

BackgroundDeep neural networks (DNNs) have the potential to revolutionize our understanding and treatment of genetic diseases. An inherent limitation of deep neural networks, however, is their high demand for data during training. To overcome this challenge, other fields, such as computer vision, use various data augmentation techniques to artificially increase the available training data for DNNs. Unfortunately, most data augmentation techniques used in other domains do not transfer well to genomic data.ResultsMost genomic data possesses peculiar properties and data augmentations may significantly alter the intrinsic properties of the data. In this work, we propose a novel data augmentation technique for genomic data inspired by biology: point mutations. By employing point mutations as substitutes for codons, we demonstrate that our newly proposed data augmentation technique enhances the performance of DNNs across various genomic tasks that involve coding regions, such as translation initiation and splice site detection.ConclusionSilent and missense mutations are found to positively influence effectiveness, while nonsense mutations and random mutations in non-coding regions generally lead to degradation. Overall, point mutation-based augmentations in genomic datasets present valuable opportunities for improving the accuracy and reliability of predictive models for DNA sequences.

Performance Evaluation of YOLOv5-based Custom Object Detection Model for Campus-Specific Scenario

This study evaluates the performance of a custom object detection model based on the YOLOv5 architecture, specifically tailored for autonomous electric vehicles. The model undergoes pre-processing using the Roboflow computer vision platform, which offers a wide range of tools for data pre-processing and model training. The experiments were conducted on a diverse dataset comprising various objects encountered in campus-specific driving scenarios, such as pedestrians, vehicles, buildings, and obstacles. The performance of the custom object detection model is assessed using standard metrics, including precision, recall, mean average precision (mAP), and intersection-over-union (IoU) at different thresholds. The training process was conducted in a controlled environment, resulting in a Precision of 0.851, a Recall of 0.831, and a mAP of 0.843. These metrics were analyzed to evaluate the YOLOv5-based custom object detection model's ability to detect and categorize objects accurately, its precision in predicting bounding boxes, and its capability to handle various object categories. We also examined the effects of different hyperparameters and data augmentation techniques on the model's performance, including variations in learning rate, batch size, and optimizer algorithms to determine their impact on accuracy and convergence. This analysis provided valuable insights into the model's strengths and weaknesses, highlighting areas for improvement and optimization. These findings are instrumental in developing and deploying advanced object detection systems to enhance the safety and reliability of autonomous electric vehicles.

Emotion Symphony: Deep Learning Journey Through Human Facial Emotions

Abstract: In today's technologically advanced world, understanding human emotions holds paramount importance in various domains ranging from human-computer interaction to mental health diagnosis. Emotions are primarily conveyed through facial expressions, making facial emotion recognition a crucial area of research. This project, "Emotion Symphony," embarks on a deep learning journey aimed at developing an efficient system for recognizing and interpreting human facial emotions. The project employs state-of-the-art deep learning techniques to extract meaningful features from facial images and map them to corresponding emotional states. Leveraging convolutional neural networks (CNNs) and recurrent neural networks (RNNs), the system is designed to analyse subtle nuances in facial expressions, capturing both spatial and temporal dependencies for accurate emotion recognition. To train the model, extensive datasets comprising diverse facial expressions are utilized, ensuring robustness and generalization of the proposed system. Additionally, data augmentation techniques are employed to enhance the model's ability to handle variations in lighting conditions, poses, and facial expressions.

Revisiting the utility of identifying nuclear grooves as unique nuclear changes by an object detector model.

Among other structures, nuclear grooves are vastly found in papillary thyroid carcinoma (PTC). Considering that the application of artificial intelligence in thyroid cytology has potential for diagnostic routine, our goal was to develop a new supervised convolutional neural network capable of identifying nuclear grooves in Diff-Quik stained whole-slide images (WSI) obtained from thyroid fineneedle aspiration. We selected 22 Diff-Quik stained cytological slides with cytological diagnosis of PTC and concordant histological diagnosis. Each of the slides was scanned, forming a WSI. Images that contained the region of interest were obtained, followed by pre-formatting, annotation of the nuclear grooves and data augmentation techniques. The final dataset was divided into training and validation groups in a 7:3 ratio. This is the first artificial intelligence model based on object detection applied to nuclear structures in thyroid cytopathology. A total of 7,255 images were obtained from 22 WSI, totaling 7,242 annotated nuclear grooves. The best model was obtained after it was submitted 15 times with the train dataset (14th epoch), with 67% true positives, 49.8% for sensitivity and 43.1% for predictive positive value. Conclusion: The model was able to develop a structure predictor rule, indicating that the application of an artificial intelligence model based on object detection in the identification of nuclear grooves is feasible. Associated with a reduction in interobserver variability and in time per slide, this demonstrates that nuclear evaluation constitutes one of the possibilities for refining the diagnosis through computational models.

Optimizing Vision Transformers for Histopathology: Pretraining and Normalization in Breast Cancer Classification

This paper introduces a self-attention Vision Transformer model specifically developed for classifying breast cancer in histology images. We examine various training strategies and configurations, including pretraining, dimension resizing, data augmentation and color normalization strategies, patch overlap, and patch size configurations, in order to evaluate their impact on the effectiveness of the histology image classification. Additionally, we provide evidence for the increase in effectiveness gathered through geometric and color data augmentation techniques. We primarily utilize the BACH dataset to train and validate our methods and models, but we also test them on two additional datasets, BRACS and AIDPATH, to verify their generalization capabilities. Our model, developed from a transformer pretrained on ImageNet, achieves an accuracy rate of 0.91 on the BACH dataset, 0.74 on the BRACS dataset, and 0.92 on the AIDPATH dataset. Using a model based on the prostate small and prostate medium HistoEncoder models, we achieve accuracy rates of 0.89 and 0.86, respectively. Our results suggest that pretraining on large-scale general datasets like ImageNet is advantageous. We also show the potential benefits of using domain-specific pretraining datasets, such as extensive histopathological image collections as in HistoEncoder, though not yet with clear advantages.

SStackGNN: Graph Data Augmentation Simplified Stacking Graph Neural Network for Twitter Bot Detection

Ensemble learning has the potential to enhance the efficacy of feeble classifiers significantly and is increasingly being utilized in Twitter bot detection. Previous methods have utilized stacking techniques to train the primary classifiers, implementing cross-validation to mitigate overfitting and enhance predictive accuracy. However, cross-validation substantially amplifies the computation time associated with stacking. To overcome this challenge, this paper presents a novel approach, the Simplified Stacking Graph Neural Network (SStackGNN), specifically designed for the detection of social bots. Our methodology leverages the power of Graph Neural Networks (GNNs) as base classifiers, enabling effective capturing of inter-account interactions. In addition, a Multilayer Perceptron (MLP) serves as a secondary classifier, amalgamating the outcomes of the foundational classifiers to generate final predictions, thereby enhancing the predictive performance of the GNNs. Instead of relying on cross-validation and distinct base classifier structures, we employ node-level, edge-level, and feature-level graph data augmentation techniques to acquire diverse foundational classifiers. This approach significantly alleviates the computational complexity while achieving superior performance. Experimental results demonstrate that our proposed SStackGNN outperforms other approaches.

AUTOMATED PNEUMONIA IDENTIFICATION THROUGH CNN-BASED ANALYSIS

Pneumonia, a condition caused by infectious agents such as bacteria, viruses, fungi, or parasites, results in lung infections and the accumulation of pus in affected tissues. Incorrect diagnosis or improper treatment can significantly impact a patient's quality of life. This project focuses on the precise identification and categorization of pneumonia-afflicted patients through the analysis of their chest X-rays. Leveraging advancements in deep learning, healthcare experts can make more accurate diagnostic decisions for various illnesses. This study presents an innovative approach utilizing Convolutional Neural Networks (CNN) to predict and distinguish between patients affected by pneumonia and those who are not, based on chest X-ray images. The methodology involves a series of convolutional and max pooling layers activated using the ReLU activation function. Subsequently, the processed data is passed through dense layers, culminating in the activation of the output neuron through the sigmoidal function. During the training process, the model exhibits increasing accuracy and decreasing loss, demonstrating its effectiveness. To prevent overfitting, data augmentation techniques are applied before model fitting. These strategies collectively enhance the robustness and reliability of the deep learning models, resulting in accurate and persuasive detection of pneumonia from chest X-ray images. Keywords: Pneumonia, CNN, VGG-16, VGG-19, Chest X-rays, Classification, Deeep learning, Data augmenatation, ReLU, softmax.

Efficient diagnosis of psoriasis and lichen planus cutaneous diseases using deep learning approach

The tendency of skin diseases to manifest in a unique and yet similar appearance, absence of enough competent dermatologists, and urgency of diagnosis and classification on time and accurately, makes the need of machine aided diagnosis blatant. This study is conducted with the purpose of broadening the research in skin disease diagnosis with computer by traversing the capabilities of deep Learning algorithms to classify two skin diseases noticeably close in appearance, Psoriasis and Lichen Planus. The resemblance between these two skin diseases is striking, often resulting in their classification within the same category. Despite this, there is a dearth of research focusing specifically on these diseases. A customized 50 layers ResNet-50 architecture of convolutional neural network is used and the results are validated through fivefold cross-validation, threefold cross-validation, and random split. By utilizing advanced data augmentation and class balancing techniques, the diversity of the dataset has increased, and the dataset imbalance has been minimized. ResNet-50 has achieved an accuracy of 89.07%, sensitivity of 86.46%, and specificity of 86.02%. With their promising results, these algorithms make the potential of machine aided diagnosis clear. Deep Learning algorithms could provide assistance to physicians and dermatologists by classification of skin diseases, with similar appearance, in real-time.

Prediction of Alzheimer's disease stages based on ResNet-Self-attention architecture with Bayesian optimization and best features selection

Alzheimer's disease (AD) is a neurodegenerative illness that impairs cognition, function, and behavior by causing irreversible damage to multiple brain areas, including the hippocampus. The suffering of the patients and their family members will be lessened with an early diagnosis of AD. The automatic diagnosis technique is widely required due to the shortage of medical experts and eases the burden of medical staff. The automatic artificial intelligence (AI)-based computerized method can help experts achieve better diagnosis accuracy and precision rates. This study proposes a new automated framework for AD stage prediction based on the ResNet-Self architecture and Fuzzy Entropy-controlled Path-Finding Algorithm (FEcPFA). A data augmentation technique has been utilized to resolve the dataset imbalance issue. In the next step, we proposed a new deep-learning model based on the self-attention module. A ResNet-50 architecture is modified and connected with a self-attention block for important information extraction. The hyperparameters were optimized using Bayesian optimization (BO) and then utilized to train the model, which was subsequently employed for feature extraction. The self-attention extracted features were optimized using the proposed FEcPFA. The best features were selected using FEcPFA and passed to the machine learning classifiers for the final classification. The experimental process utilized a publicly available MRI dataset and achieved an improved accuracy of 99.9%. The results were compared with state-of-the-art (SOTA) techniques, demonstrating the improvement of the proposed framework in terms of accuracy and time efficiency.

Optimizing Cattle Behavior Analysis in Precision Livestock Farming: Integrating YOLOv7-E6E with AutoAugment and GridMask to Enhance Detection Accuracy

Recently, the growing demand for meat has increased interest in precision livestock farming (PLF), wherein monitoring livestock behavior is crucial for assessing animal health. We introduce a novel cattle behavior detection model that leverages data from 2D RGB cameras. It primarily employs you only look once (YOLO)v7-E6E, which is a real-time object detection framework renowned for its efficiency across various applications. Notably, the proposed model enhances network performance without incurring additional inference costs. We primarily focused on performance enhancement and evaluation of the model by integrating AutoAugment and GridMask to augment the original dataset. AutoAugment, a reinforcement learning algorithm, was employed to determine the most effective data augmentation policy. Concurrently, we applied GridMask, a novel data augmentation technique that systematically eliminates square regions in a grid pattern to improve model robustness. Our results revealed that when trained on the original dataset, the model achieved a mean average precision (mAP) of 88.2%, which increased by 2.9% after applying AutoAugment. The performance was further improved by combining AutoAugment and GridMask, resulting in a notable 4.8% increase in the mAP, thereby achieving a final mAP of 93.0%. This demonstrates the efficacy of these augmentation strategies in improving cattle behavior detection for PLF.

An Improved CNN model for Identifying Tomato Leaf Diseases

A major loss in gross domestic product, quantity and quality of products produced, as well as tomato production, is caused by diseases in tomato leaves due to  which farmers have a difficult time in controlling and monitoring the health of tomato leaves, one of which is leaf disease. In our  project, we developed an Enhanced CNN by using data augmentation techniques to identify the seven classes(blight ,leaf curl, leaf miner ,Alteneria, leaf spot,  cutwork infected) of  tomato leaf diseases. Using 27807 trainable parameters, the enhanced CNN obtains the maximum training accuracy of ninety nine point nine eight percent(99.98%) and validating accuracy of ninety eight point four percent(98.4%) .With fewer parameters, The Enhanced CNN can more accurately determine the type of illness of tomato leaf. Our Enhanced CNN model also determine the type of illness of tomato leaf when tested with the images of diseased tomato leaf collected from the internet sources.Using 152850 trainable parameters the enhanced CNN obtain the maximum training accuracy of 99.68% and validation accuracy of 89%.

Empowering modern E-commerce consumers: Optimized deep learning for product image classification in home shopping

Accurately classifying products in e-commerce is critical for enhancing user experience, but it remains challenging due to data quality issues and the dynamic nature of product categories. Customers are increasingly relying on visual information to make informed purchasing decisions, emphasizing the importance of accurate product classification using images. In this paper, an innovative approach called SSWSO_LeNet is proposed for product image classification in e-commerce. The method involves preprocessing the input images using Region of Interest (RoI) and Adaptive Wiener Filters to improve image quality and reduce unwanted distortions. Data augmentation techniques are then applied to increase the diversity of the dataset and the robustness of the model. To address this, we propose SSWSO_LeNet, integrating Squirrel Search Algorithm (SSA) and War Strategy Optimization (WSO) with LeNet. SSA mimics southern flying squirrels’ foraging behavior to find global optima efficiently, while WSO balances exploration and exploitation stages, enhancing classification accuracy. Experimental results show SSWSO_LeNet outperforms state-of-the-art models with an impressive accuracy of 0.976, sensitivity of 0.877, and specificity of 0.857. By leveraging SSA, WSO, and LeNet, SSWSO_LeNet not only improves classification accuracy but also reduces reliance on human editors, decreasing both cost and time in e-commerce product classification.

Detection and identification of plant leaf diseases using YOLOv4.

Detecting plant leaf diseases accurately and promptly is essential for reducing economic consequences and maximizing crop yield. However, farmers' dependence on conventional manual techniques presents a difficulty in accurately pinpointing particular diseases. This research investigates the utilization of the YOLOv4 algorithm for detecting and identifying plant leaf diseases. This study uses the comprehensive Plant Village Dataset, which includes over fifty thousand photos of healthy and diseased plant leaves from fourteen different species, to develop advanced disease prediction systems in agriculture. Data augmentation techniques including histogram equalization and horizontal flip were used to improve the dataset and strengthen the model's resilience. A comprehensive assessment of the YOLOv4 algorithm was conducted, which involved comparing its performance with established target identification methods including Densenet, Alexanet, and neural networks. When YOLOv4 was used on the Plant Village dataset, it achieved an impressive accuracy of 99.99%. The evaluation criteria, including accuracy, precision, recall, and f1-score, consistently showed high performance with a value of 0.99, confirming the effectiveness of the proposed methodology. This study's results demonstrate substantial advancements in plant disease detection and underscore the capabilities of YOLOv4 as a sophisticated tool for accurate disease prediction. These developments have significant significance for everyone involved in agriculture, researchers, and farmers, providing improved capacities for disease control and crop protection.

Oil spill detection and classification through deep learning and tailored data augmentation

Oil spill detection is an important task for protecting and minimizing the harmful effects of oil on the marine ecosystem. Currently, the application of images from unmanned aerial vehicles, along with deep learning, is widely employed. Although these methods have yielded good results, the issue of oil spill classification based on these methods has not received much attention at present. In this research, a deep learning model with a dual attention mechanism consisting of two modules was utilized. The first module focuses on capturing the spatial relationships between each pixel and the entire image, the second module identifies the characteristics between channels in the image, thereby enhancing the ability to detect and classify oil. Additionally, a data augmentation technique based on the Generative Adversarial Networks model was refined and employed to improve the model's accuracy. Experimental results, obtained through comparisons between dataset construction methods, the use of different encoders and decoders, and adjustments hyperparameters, reveal that the best model achieves a mean Intersection over Union by 72.49%. Data augmentation techniques also contribute to a 2.56% increase in mean Intersection over Union. The findings of this research provide a feasible solution not only for detecting but also for classifying oil spills, thereby assisting marine environmental managers in making timely decisions to respond to oil spill accidents.

Development of generic metabolic Raman calibration models using solution titration in aqueous phase and data augmentation for in-line cell culture analysis.

This study presents a novel approach for developing generic metabolic Raman calibration models for in-line cell culture analysis using glucose and lactate stock solution titration in an aqueous phase and data augmentation techniques. First, a successful set-up of the titration method was achieved by adding glucose or lactate solution at several different constant rates into the aqueous phase of a bench-top bioreactor. Subsequently, the in-line glucose and lactate concentration were calculated and interpolated based on the rate of glucose and lactate addition, enabling data augmentation and enhancing the robustness of the metabolic calibration model. Nine different combinations of spectra pretreatment, wavenumber range selection, and number of latent variables were evaluated and optimized using aqueous titration data as training set and a historical cell culture data set as validation and prediction set. Finally, Raman spectroscopy data collected from 11 historical cell culture batches (spanning four culture modes and scales ranging from 3to 200 L) were utilized to predict the corresponding glucose and lactate values. The results demonstrated a high prediction accuracy, with an average root mean square errors of predictionof 0.65 g/L for glucose, and 0.48 g/L for lactate. This innovative method establishes a generic metabolic calibration model, and its applicability can be extended to other metabolites, reducing the cost of deploying real-time cell culture monitoring using Raman spectroscopy in bioprocesses.

Classification of complex local environments in systems of particle shapes through shape symmetry-encoded data augmentation.

Detecting and analyzing the local environment is crucial for investigating the dynamical processes of crystal nucleation and shape colloidal particle self-assembly. Recent developments in machine learning provide a promising avenue for better order parameters in complex systems that are challenging to study using traditional approaches. However, the application of machine learning to self-assembly on systems of particle shapes is still underexplored. To address this gap, we propose a simple, physics-agnostic, yet powerful approach that involves training a multilayer perceptron (MLP) as a local environment classifier for systems of particle shapes, using input features such as particle distances and orientations. Our MLP classifier is trained in a supervised manner with a shape symmetry-encoded data augmentation technique without the need for any conventional roto-translations invariant symmetry functions. We evaluate the performance of our classifiers on four different scenarios involving self-assembly of cubic structures, two-dimensional and three-dimensional patchy particle shape systems, hexagonal bipyramids with varying aspect ratios, and truncated shapes with different degrees of truncation. The proposed training process and data augmentation technique are both straightforward and flexible, enabling easy application of the classifier to other processes involving particle orientations. Our work thus presents a valuable tool for investigating self-assembly processes on systems of particle shapes, with potential applications in structure identification of any particle-based or molecular system where orientations can be defined.