Search results

Create and share a MATLAB library that performs data augmentation algorithms for audio data. This study aims to help machine learning researchers to improve their models using the algorithms proposed by the authors.

The authors structured our library into methods to augment raw audio data and spectrograms. In the paper, the authors describe the structure of the library and give a brief explanation of how every function works. The authors then perform experiments to show that the library is effective.

The authors prove that the library is efficient using a competitive dataset. The authors try multiple data augmentation approaches proposed by them and show that they improve the performance.

A MATLAB library specifically designed for data augmentation was not available before. The authors are the first to provide an efficient and parallel implementation of a large number of algorithms.

In recent years, the convolutional neural network (CNN) based deep learning approach has succeeded in data-mining the relationship between microstructures and macroscopic properties of materials. However, such CNN models usually rely heavily on a large set of labeled images to ensure the accuracy and generalization ability of the predictive models. Unfortunately, in many fields, acquiring image data is expensive and inconvenient. This study aims to propose a data augmentation technique to enhance the performance of the CNN models for linking microstructural images to the macroscopic properties of composites.

Microstructures of composites are synthesized using discrete element simulations and Potts kinetic Monte Carlo simulations. Macroscopic properties such as the elastic modulus, Poisson's ratio, shear modulus, coefficient of thermal expansion, and triple-phase boundary length density are extracted on representative volume elements. The CNN model is trained using the 3D microstructural images as inputs and corresponding macroscopic properties as the labels. The comparison of the predictive performance of the CNN models with and without data augmentation treatment are compared.

The comparison between the prediction performance of CNN models with and without data augmentation showed that the former reduced the weighted mean absolute percentage error (WMAPE) for the prediction from 5.1627% to 1.7014%. This significant reduction signifies that the proposed data augmentation method can effectively enhance the generalization ability and robustness of CNN models.

This study demonstrates that data augmentation is beneficial for solving the problems of model overfitting, data scarcity, and sample imbalance for CNN-based deep learning tasks at a low cost. By developing more and advanced data augmentation techniques, deep learning accelerated homogenization will boost the multi-scale computational mechanics and materials.

This paper aims to propose a new deep learning technique to detect the type of material to improve automated construction quality monitoring.

A new data augmentation approach that has improved the model robustness against different illumination conditions and overfitting is proposed. This study uses data augmentation at test time and adds outlier samples to training set to prevent over-fitted network training. For data augmentation at test time, five segments are extracted from each sample image and fed to the network. For these images, the network outputting average values is used as the final prediction. Then, the proposed approach is evaluated on multiple deep networks used as material classifiers. The fully connected layers are removed from the end of the networks, and only convolutional layers are retained.

The proposed method is evaluated on recognizing 11 types of building materials which include 1,231 images taken from several construction sites. Each image resolution is 4,000 × 3,000. The images are captured with different illumination and camera positions. Different illumination conditions lead to trained networks that are more robust against various environmental conditions. Using VGG16 model, an accuracy of 97.35% is achieved outperforming existing approaches.

It is believed that the proposed method presents a new and robust tool for detecting and classifying different material types. The automated detection of material will aid to monitor the quality and see whether the right type of material has been used in the project based on contract specifications. In addition, the proposed model can be used as a guideline for performing quality control (QC) in construction projects based on project quality plan. It can also be used as an input for automated progress monitoring because the material type detection will provide a critical input for object detection.

Several studies have been conducted to perform quality management, but there are some issues that need to be addressed. In most previous studies, a very limited number of material types were examined. In addition, although some studies have reported high accuracy to detect material types (Bunrit et al., 2020), their accuracy is dramatically reduced when they are used to detect materials with similar texture and color. In this research, the authors propose a new method to solve the mentioned shortcomings.

This paper aims to identify data set entities in scientific literature. To address poor recognition caused by a lack of training corpora in existing studies, a distant supervised learning-based approach is proposed to identify data set entities automatically from large-scale scientific literature in an open domain.

Firstly, the authors use a dictionary combined with a bootstrapping strategy to create a labelled corpus to apply supervised learning. Secondly, a bidirectional encoder representation from transformers (BERT)-based neural model was applied to identify data set entities in the scientific literature automatically. Finally, two data augmentation techniques, entity replacement and entity masking, were introduced to enhance the model generalisability and improve the recognition of data set entities.

In the absence of training data, the proposed method can effectively identify data set entities in large-scale scientific papers. The BERT-based vectorised representation and data augmentation techniques enable significant improvements in the generality and robustness of named entity recognition models, especially in long-tailed data set entity recognition.

This paper provides a practical research method for automatically recognising data set entities in scientific literature. To the best of the authors’ knowledge, this is the first attempt to apply distant learning to the study of data set entity recognition. The authors introduce a robust vectorised representation and two data augmentation strategies (entity replacement and entity masking) to address the problem inherent in distant supervised learning methods, which the existing research has mostly ignored. The experimental results demonstrate that our approach effectively improves the recognition of data set entities, especially long-tailed data set entities.

Construction action recognition is essential to efficiently manage productivity, health and safety risks. These can be achieved by tracking and monitoring construction work. This study aims to examine the performance of a variant of deep convolutional neural networks (CNNs) for recognizing actions of construction workers from images of signals of time-series data.

This paper adopts Inception v1 to classify actions involved in carpentry and painting activities from images of motion data. Augmented time-series data from wearable sensors attached to worker's lower arms are converted to signal images to train an Inception v1 network. Performance of Inception v1 is compared with the highest performing supervised learning classifier, k-nearest neighbor (KNN).

Results show that the performance of Inception v1 network improved when trained with signal images of the augmented data but at a high computational cost. Inception v1 network and KNN achieved an accuracy of 95.2% and 99.8%, respectively when trained with 50-fold augmented carpentry dataset. The accuracy of Inception v1 and KNN with 10-fold painting augmented dataset is 95.3% and 97.1%, respectively.

Only acceleration data of the lower arm of the two trades were used for action recognition. Each signal image comprises 20 datasets.

Little has been reported on recognizing construction workers' actions from signal images. This study adds value to the existing literature, in particular by providing insights into the extent to which a deep CNN can classify subtasks from patterns in signal images compared to a traditional best performing shallow network.

The stock indexes are an important issue for investors, and in this paper good trading strategies will be aimed to be adopted according to the accurate prediction of the stock indexes to chase high returns.

To avoid the problem of insufficient financial data for daily stock indexes prediction during modeling, a data augmentation method based on time scale transformation (DATT) was introduced. After that, a new deep learning model which combined DATT and NGRU (DATT-nested gated recurrent units (NGRU)) was proposed for stock indexes prediction. The proposed models and their competitive models were used to test the stock indexes prediction and simulated trading in five stock markets of China and the United States.

The experimental results demonstrated that both NGRU and DATT-NGRU outperformed the other recurrent neural network (RNN) models in the daily stock indexes prediction.

A novel RNN with NGRU and data augmentation is proposed. It uses the nested structure to increase the depth of the deep learning model.

Press-fit with force and displacement monitoring is commonly adopted in automotive mechatronic system assembling. However, suitable methods for the press-fit study are still at initial investigation phase. The sequential data physical meaning, small data sets from different resources and computing efficiency should be considered. Therefore, this paper aims to better identify press-fit fault types.

This paper proposed one-dimensional convolutional neural network (1DCNN)–long short-term memory (LSTM) method to perform press-fit fault diagnosis into automotive assembling practice which is in accordance with current product development procedure. Specialized data augmentation method is proposed to merge different data resources and increase the sample size. Referring one-way sequential data characteristics, LSTM and batch normalization layers are integrated in 1DCNN to improve the performance.

The proposed 1DCNN-LSTM method is feasible with small data sets from different sources. Using data augmentation to make data unified and sample size increased, the accuracy could reach more than 99%. Training time has reduced from 90 s/Epoch to 4 s/Epoch compare to pure LSTM method.

The proposed method shows better performance with less training time compared to LSTM. Therefore, the method has practical value and is worthy of industrial application.

Automation of detecting cracked surfaces on buildings or in any industrially manufactured products is emerging nowadays. Detection of the cracked surface is a challenging task for inspectors. Image-based automatic inspection of cracks can be very effective when compared to human eye inspection. With the advancement in deep learning techniques, by utilizing these methods the authors can create automation of work in a particular sector of various industries.

In this study, an upgraded convolutional neural network-based crack detection method has been proposed. The dataset consists of 3,886 images which include cracked and non-cracked images. Further, these data have been split into training and validation data. To inspect the cracks more accurately, data augmentation was performed on the dataset, and regularization techniques have been utilized to reduce the overfitting problems. In this work, VGG19, Xception and Inception V3, along with Resnet50 V2 CNN architectures to train the data.

A comparison between the trained models has been performed and from the obtained results, Xception performs better than other algorithms with 99.54% test accuracy. The results show detecting cracked regions and firm non-cracked regions is very efficient by the Xception algorithm.

The proposed method can be way better back to an automatic inspection of cracks in buildings with different design patterns such as decorated historical monuments.

Because of the fast-growing digital image collections on online platforms and the transfer learning ability of deep learning technology, image classification could be improved and implemented for the hostel domain, which has complex clusters of image contents. This paper aims to test the potential of 11 pretrained convolutional neural network (CNN) with transfer learning for hostel image classification on the first hostel image database to advance the knowledge and fill the gap academically, as well as to suggest an alternative solution in optimal image classification with less labour cost and human errors to those who manage hostel image collections.

The hostel image database is first created with data pre-processing steps, data selection and data augmentation. Then, the systematic and comprehensive investigation is divided into seven experiments to test 11 pretrained CNNs which transfer learning was applied and parameters were fine-tuned to match this newly created hostel image dataset. All experiments were conducted in Google Colaboratory environment using PyTorch.

The 7,350 hostel image database is created and labelled into seven classes. Furthermore, its experiment results highlight that DenseNet 121 and DenseNet 201 have the greatest potential for hostel image classification as they outperform other CNNs in terms of accuracy and training time.

The fact that there is no existing academic work dedicating to test pretrained CNNs with transfer learning for hostel image classification and no existing hostel image-only database have made this paper a novel contribution.

Computed tomography (CT) scan can provide valuable information in the diagnosis of lung diseases. To detect the location of the cancerous lung nodules, this work uses novel deep learning methods. The majority of the early investigations used CT, magnetic resonance and mammography imaging. Using appropriate procedures, the professional doctor in this sector analyses these images to discover and diagnose the various degrees of lung cancer. All of the methods used to discover and detect cancer illnesses are time-consuming, expensive and stressful for the patients. To address all of these issues, appropriate deep learning approaches for analyzing these medical images, which included CT scan images, were utilized.

Radiologists currently employ chest CT scans to detect lung cancer at an early stage. In certain situations, radiologists' perception plays a critical role in identifying lung melanoma which is incorrectly detected. Deep learning is a new, capable and influential approach for predicting medical images. In this paper, the authors employed deep transfer learning algorithms for intelligent classification of lung nodules. Convolutional neural networks (VGG16, VGG19, MobileNet and DenseNet169) are used to constrain the input and output layers of a chest CT scan image dataset.

The collection includes normal chest CT scan pictures as well as images from two kinds of lung cancer, squamous and adenocarcinoma impacted chest CT scan images. According to the confusion matrix results, the VGG16 transfer learning technique has the highest accuracy in lung cancer classification with 91.28% accuracy, followed by VGG19 with 89.39%, MobileNet with 85.60% and DenseNet169 with 83.71% accuracy, which is analyzed using Google Collaborator.

The proposed approach using VGG16 maximizes the classification accuracy when compared to VGG19, MobileNet and DenseNet169. The results are validated by computing the confusion matrix for each network type.