Search results

This paper aims to evaluate and compare the performance of different computer vision algorithms in the context of visual servoing for augmented robot perception and autonomy.

The authors evaluated and compared three different approaches: a feature-based approach, a hybrid approach and a machine-learning-based approach. To evaluate the performance of the approaches, experiments were conducted in a simulated environment using the PyBullet physics simulator. The experiments included different levels of complexity, including different numbers of distractors, varying lighting conditions and highly varied object geometry.

The experimental results showed that the machine-learning-based approach outperformed the other two approaches in terms of accuracy and robustness. The approach could detect and locate objects in complex scenes with high accuracy, even in the presence of distractors and varying lighting conditions. The hybrid approach showed promising results but was less robust to changes in lighting and object appearance. The feature-based approach performed well in simple scenes but struggled in more complex ones.

This paper sheds light on the superiority of a hybrid algorithm that incorporates a deep neural network in a feature detector for image-based visual servoing, which demonstrates stronger robustness in object detection and location against distractors and lighting conditions.

COVID-19 has occurred in more than 150 countries and causes a huge impact on the health of many people. The main purpose of this work is, COVID-19 has occurred in more than 150 countries and causes a huge impact on the health of many people. The COVID-19 diagnosis is required to detect at the beginning stage and special attention should be given to them. The fastest way to detect the COVID-19 infected patients is detecting through radiology and radiography images. The few early studies describe the particular abnormalities of the infected patients in the chest radiograms. Even though some of the challenges occur in concluding the viral infection traces in X-ray images, the convolutional neural network (CNN) can determine the patterns of data between the normal and infected X-rays that increase the detection rate. Therefore, the researchers are focusing on developing a deep learning-based detection model.

The main intention of this proposal is to develop the enhanced lung segmentation and classification of diagnosing the COVID-19. The main processes of the proposed model are image pre-processing, lung segmentation and deep classification. Initially, the image enhancement is performed by contrast enhancement and filtering approaches. Once the image is pre-processed, the optimal lung segmentation is done by the adaptive fuzzy-based region growing (AFRG) technique, in which the constant function for fusion is optimized by the modified deer hunting optimization algorithm (M-DHOA). Further, a well-performing deep learning algorithm termed adaptive CNN (A-CNN) is adopted for performing the classification, in which the hidden neurons are tuned by the proposed DHOA to enhance the detection accuracy. The simulation results illustrate that the proposed model has more possibilities to increase the COVID-19 testing methods on the publicly available data sets.

From the experimental analysis, the accuracy of the proposed M-DHOA–CNN was 5.84%, 5.23%, 6.25% and 8.33% superior to recurrent neural network, neural networks, support vector machine and K-nearest neighbor, respectively. Thus, the segmentation and classification performance of the developed COVID-19 diagnosis by AFRG and A-CNN has outperformed the existing techniques.

This paper adopts the latest optimization algorithm called M-DHOA to improve the performance of lung segmentation and classification in COVID-19 diagnosis using adaptive K-means with region growing fusion and A-CNN. To the best of the authors’ knowledge, this is the first work that uses M-DHOA for improved segmentation and classification steps for increasing the convergence rate of diagnosis.

In the process of continuous casting and rolling of steel plate, due to the influence of rolling equipment and process, there are scratches, inclusions, patches, scabs and pitted surfaces on the surface of steel plate, which will not only affect the corrosion resistance, wear resistance and fatigue strength of steel plate but also may cause production accidents. Therefore, the detection of steel plate surface defect must be strengthened to ensure the production quality of steel plate and the smooth development of industrial construction.

(1) A steel plate surface defect detection technology based on small datasets is proposed, which can detect multiple surface defects and fill in the blank of scab defect detection. (2) A detection system based on intelligent recognition technology is built. The steel plate images are collected by the front-end monitoring device, then transmitted to the back-end monitoring center and processed by the embedded intelligent algorithms. (3) In order to reduce the impact of external light on the image, an improved Multi-Scale Retinex (MSR) enhancement algorithm based on adaptive weight calculation is proposed, which lays the foundation for subsequent object segmentation and feature extraction. (4) According to the different factors such as the cause and shape, the texture and shape features are combined to classify different defects on the steel plate surface. The defect classification model is constructed and the classification results are recorded and stored, which has certain application value in the field of steel plate surface defect detection. (5) The practicability and effectiveness of the proposed method are verified by comparison with other methods, and the field running tests are conducted based on the equipment commissioning field of China Heavy Machinery Institute.

When applied to small dataset, the precision of the proposed method is 94.5% and the time is 23.7 ms. In order to compare with deep learning technology, after expanding the image dataset, the precision and detection time of this paper are 0.948 and 24.2 ms, respectively. The proposed method is superior to other traditional image processing and deep learning methods. And the field recognition precision is 91.7%.

In brief, the steel plate surface defect detection technology based on computer vision is effective, but the previous attempts and methods are not comprehensive and the accuracy and detection speed need to be improved. Therefore, a more practical and comprehensive technology is developed in this paper. The main contributions are as follows: (1) A steel plate surface defect detection technology based on small datasets is proposed, which can detect multiple surface defects and fill in the blank of scab defect detection. (2) A detection system based on intelligent recognition technology is built. The steel plate images are collected by the front-end monitoring device, then transmitted to the back-end monitoring center and processed by the embedded intelligent algorithms. (3) In order to reduce the impact of external light on the image, an improved MSR enhancement algorithm based on adaptive weight calculation is proposed, which lays the foundation for subsequent object segmentation and feature extraction. (4) According to the different factors such as the cause and shape, the texture and shape features are combined to classify different defects on the steel plate surface. The defect classification model is constructed and the classification results are recorded and stored, which has certain application value in the field of steel plate surface defect detection. (5) The practicability and effectiveness of the proposed method are verified by comparison with other methods, and the field running tests are conducted based on the equipment commissioning field of China Heavy Machinery Institute.

An accurate prediction of the deformation of retaining structures is critical for ensuring the stability and safety of braced deep excavations, while the high nonlinear and complex nature of the deformation makes the prediction challenging. This paper proposes an explainable boosted combining global and local feature multivariate regression (EB-GLFMR) model with high accuracy, robustness and interpretability to predict the deformation of retaining structures during braced deep excavations.

During the model development, the time series of deformation data is decomposed using a locally weighted scatterplot smoothing technique into trend and residual terms. The trend terms are analyzed through multiple adaptive spline regressions. The residual terms are reconstructed in phase space to extract both global and local features, which are then fed into a gradient-boosting model for prediction.

The proposed model outperforms other established approaches in terms of accuracy and robustness, as demonstrated through analyzing two cases of braced deep excavations.

The model is designed for the prediction of the deformation of deep excavations with stepped, chaotic and fluctuating features. Further research needs to be conducted to expand the model applicability to other time series deformation data.

The model provides an efficient, robust and transparent approach to predict deformation during braced deep excavations. It serves as an effective decision support tool for engineers to ensure the stability and safety of deep excavations.

The model captures the global and local features of time series deformation of retaining structures and provides explicit expressions and feature importance for deformation trends and residuals, making it an efficient and transparent approach for deformation prediction.

For achieving the profitable human activity recognition (HAR) method, this paper solves the HAR problem under wireless body area network (WBAN) using a developed ensemble learning approach. The purpose of this study is,to solve the HAR problem under WBAN using a developed ensemble learning approach for achieving the profitable HAR method. There are three data sets used for this HAR in WBAN, namely, human activity recognition using smartphones, wireless sensor data mining and Kaggle. The proposed model undergoes four phases, namely, “pre-processing, feature extraction, feature selection and classification.” Here, the data can be preprocessed by artifacts removal and median filtering techniques. Then, the features are extracted by techniques such as “t-Distributed Stochastic Neighbor Embedding”, “Short-time Fourier transform” and statistical approaches. The weighted optimal feature selection is considered as the next step for selecting the important features based on computing the data variance of each class. This new feature selection is achieved by the hybrid coyote Jaya optimization (HCJO). Finally, the meta-heuristic-based ensemble learning approach is used as a new recognition approach with three classifiers, namely, “support vector machine (SVM), deep neural network (DNN) and fuzzy classifiers.” Experimental analysis is performed.

The proposed HCJO algorithm was developed for optimizing the membership function of fuzzy, iteration limit of SVM and hidden neuron count of DNN for getting superior classified outcomes and to enhance the performance of ensemble classification.

The accuracy for enhanced HAR model was pretty high in comparison to conventional models, i.e. higher than 6.66% to fuzzy, 4.34% to DNN, 4.34% to SVM, 7.86% to ensemble and 6.66% to Improved Sealion optimization algorithm-Attention Pyramid-Convolutional Neural Network-AP-CNN, respectively.

The suggested HAR model with WBAN using HCJO algorithm is accurate and improves the effectiveness of the recognition.

This study updates the literature review of Jones (1987) published in this journal. The study pays particular attention to two important themes that have shaped the field over the past 35 years: (1) the development of a range of innovative new statistical learning methods, particularly advanced machine learning methods such as stochastic gradient boosting, adaptive boosting, random forests and deep learning, and (2) the emergence of a wide variety of bankruptcy predictor variables extending beyond traditional financial ratios, including market-based variables, earnings management proxies, auditor going concern opinions (GCOs) and corporate governance attributes. Several directions for future research are discussed.

This study provides a systematic review of the corporate failure literature over the past 35 years with a particular focus on the emergence of new statistical learning methodologies and predictor variables. This synthesis of the literature evaluates the strength and limitations of different modelling approaches under different circumstances and provides an overall evaluation the relative contribution of alternative predictor variables. The study aims to provide a transparent, reproducible and interpretable review of the literature. The literature review also takes a theme-centric rather than author-centric approach and focuses on structured themes that have dominated the literature since 1987.

There are several major findings of this study. First, advanced machine learning methods appear to have the most promise for future firm failure research. Not only do these methods predict significantly better than conventional models, but they also possess many appealing statistical properties. Second, there are now a much wider range of variables being used to model and predict firm failure. However, the literature needs to be interpreted with some caution given the many mixed findings. Finally, there are still a number of unresolved methodological issues arising from the Jones (1987) study that still requiring research attention.

The study explains the connections and derivations between a wide range of firm failure models, from simpler linear models to advanced machine learning methods such as gradient boosting, random forests, adaptive boosting and deep learning. The paper highlights the most promising models for future research, particularly in terms of their predictive power, underlying statistical properties and issues of practical implementation. The study also draws together an extensive literature on alternative predictor variables and provides insights into the role and behaviour of alternative predictor variables in firm failure research.

This paper aims to present a comprehensive overview of the patterns and trends of publications on artificial intelligence (AI) in personalised learning. It addresses the need to investigate the intellectual structure and development of this area in view of the growing amount of related research and practices.

A bibliometric analysis was conducted to cover publications on AI in personalised learning published from 2000 to 2022, including a total of 1,005 publications collected from the Web of Science and Scopus. The patterns and trends in terms of sources of publications, intellectual structure and major topics were analysed.

Research on AI in personalised learning has been widely published in various sources. The intellectual bases of related work were mostly on studies on the application of AI technologies in education and personalised learning. The relevant research covered mainly AI technologies and techniques, as well as the design and development of AI systems to support personalised learning. The emerging topics have addressed areas such as big data, learning analytics and deep learning.

This study depicted the research hotspots of personalisation in learning with the support of AI and illustrated the evolution and emerging trends in the field. The results highlight its latest developments and the need for future work on diverse means to support personalised learning with AI, the pedagogical issues, as well as teachers’ roles and teaching strategies.

Recently, the convolutional neural network (ConvNet) has a wide application in the classification of motor imagery EEG signals. However, the low signal-to-noise electroencephalogram (EEG) signals are collected under the interference of noises. However, the conventional ConvNet model cannot directly solve this problem. This study aims to discuss the aforementioned issues.

To solve this problem, this paper adopted a novel residual shrinkage block (RSB) to construct the ConvNet model (RSBConvNet). During the feature extraction from EEG signals, the proposed RSBConvNet prevented the noise component in EEG signals, and improved the classification accuracy of motor imagery. In the construction of RSBConvNet, the author applied the soft thresholding strategy to prevent the non-related motor imagery features in EEG signals. The soft thresholding was inserted into the residual block (RB), and the suitable threshold for the current EEG signals distribution can be learned by minimizing the loss function. Therefore, during the feature extraction of motor imagery, the proposed RSBConvNet de-noised the EEG signals and improved the discriminative of classification features.

Comparative experiments and ablation studies were done on two public benchmark datasets. Compared with conventional ConvNet models, the proposed RSBConvNet model has obvious improvements in motor imagery classification accuracy and Kappa coefficient. Ablation studies have also shown the de-noised abilities of the RSBConvNet model. Moreover, different parameters and computational methods of the RSBConvNet model have been tested on the classification of motor imagery.

Based on the experimental results, the RSBConvNet constructed in this paper has an excellent recognition accuracy of MI-BCI, which can be used for further applications for the online MI-BCI.

The purpose of this paper is to propose a whole set of methods for underwater target detection, because most underwater objects have small samples, low quality underwater images problems such as detail loss, low contrast and color distortion, and verify the feasibility of the proposed methods through experiments.

The improved RGHS algorithm to enhance the original underwater target image is proposed, and then the YOLOv4 deep learning network for underwater small sample targets detection is improved based on the combination of traditional data expansion method and Mosaic algorithm, expanding the feature extraction capability with SPP (Spatial Pyramid Pooling) module after each feature extraction layer to extract richer feature information.

The experimental results, using the official dataset, reveal a 3.5% increase in average detection accuracy for three types of underwater biological targets compared to the traditional YOLOv4 algorithm. In underwater robot application testing, the proposed method achieves an impressive 94.73% average detection accuracy for the three types of underwater biological targets.

Underwater target detection is an important task for underwater robot application. However, most underwater targets have the characteristics of small samples, and the detection of small sample targets is a comprehensive problem because it is affected by the quality of underwater images. This paper provides a whole set of methods to solve the problems, which is of great significance to the application of underwater robot.

Affective states in learning have gained immense attention in education. The precise affective-states prediction can increase the learning gain by adapting targeted interventions that can adjust the changes in individual affective states of students. Several techniques are devised for predicting the affective states considering audio, video and biosensors. Still, the system that relies on analyzing audio and video cannot certify anonymity and is subjected to privacy problems.

A new strategy, termed rider squirrel search algorithm-based deep long short-term memory (RiderSSA-based deep LSTM) is devised for affective-state prediction. The deep LSTM training is done by the proposed RiderSSA. Here, RiderSSA-based deep LSTM effectively predicts the affective states like confusion, engagement, frustration, anger, happiness, disgust, boredom, surprise and so on. In addition, the learning styles are predicted based on the extracted features using rider neural network (RideNN), for which the Felder–Silverman learning-style model (FSLSM) is considered. Here, the RideNN classifies the learners. Finally, the course ID, student ID, affective state, learning style, exam score and course completion are taken as output data to determine the correlative study.

The proposed RiderSSA-based deep LSTM provided enhanced efficiency with elevated accuracy of 0.962 and the highest correlation of 0.406.

The proposed method based on affective prediction obtained maximal accuracy and the highest correlation. Thus, the method can be applied to the course recommendation system based on affect prediction.