Accuracy Of CNN Research Articles

An integrated method for detecting lung cancer via CT scanning via optimization, deep learning, and IoT data transmission.

With its increasing global prevalence, lung cancer remains a critical health concern. Despite the advancement of screening programs, patient selection and risk stratification pose significant challenges. This study addresses the pressing need for early detection through a novel diagnostic approach that leverages innovative image processing techniques. The urgency of early lung cancer detection is emphasized by its alarming growth worldwide. While computed tomography (CT) surpasses traditional X-ray methods, a comprehensive diagnosis requires a combination of imaging techniques. This research introduces an advanced diagnostic tool implemented through image processing methodologies. The methodology commences with histogram equalization, a crucial step in artifact removal from CT images sourced from a medical database. Accurate lung CT image segmentation, which is vital for cancer diagnosis, follows. The Otsu thresholding method and optimization, employing Colliding Bodies Optimization (CBO), enhance the precision of the segmentation process. A local binary pattern (LBP) is deployed for feature extraction, enabling the identification of nodule sizes and precise locations. The resulting image underwent classification using the densely connected CNN (DenseNet) deep learning algorithm, which effectively distinguished between benign and malignant tumors. The proposed CBO+DenseNet CNN exhibits remarkable performance improvements over traditional methods. Notable enhancements in accuracy (98.17%), specificity (97.32%), precision (97.46%), and recall (97.89%) are observed, as evidenced by the results from the fractional randomized voting model (FRVM). These findings highlight the potential of the proposed model as an advanced diagnostic tool. Its improved metrics promise heightened accuracy in tumor classification and localization. The proposed model uniquely combines Colliding Bodies Optimization (CBO) with DenseNet CNN, enhancing segmentation and classification accuracy for lung cancer detection, setting it apart from traditional methods with superior performance metrics.

An optimized deep hybrid learning for multi-channel EEG-based driver drowsiness detection

Driver drowsiness is a severe issue that has contributed to numerous fatal accidents and injuries. Thus, detecting driver drowsiness is an important task that has been the subject of intensive research in recent years. There have been numerous proposed physiological signals for detecting driver drowsiness. However, the Electroencephalographic (EEG) signal is the most commonly used due to its direct relationship with drowsiness and its simplicity of acquisition. Recently, different Machine Learning (ML) and Deep Learning (DL) models have been proposed to detect driver drowsiness. This study utilized a publicly accessible dataset containing twelve healthy participants. Reading numerous research papers, we determined no specific EEG-based drowsiness preprocessing parameter values. Consequently, as a first step, and for the first time, to our knowledge in this field, we applied an optimization algorithm to determine the optimal preprocessing parameter values using a CNN model and accuracy as the objective function. The obtained results demonstrated the importance of selecting the correct values, as the mean accuracy score increased from 91% before optimization to 95% after optimization for the proposed CNN. The training time has been reduced. Also, as a second step, we have used the Optuna Hyperparameter optimization framework to select the optimal CNN Hyperparameters, which increased the mean accuracy from 95% to 97%. Finally, to take advantage of Deep Hybrid learning, we have fused the CNN “features extractor” with seven ML classifiers, with the CNN-SVM classifier achieving the highest average accuracy of 99.9%, and the training time has been reduced to a shallow value.

Sentiment Analysis of FLO Applications For Women's Needs Using The CNN And LSTM Algorithms.

One application that helps women is an application that can track menstrual schedules, plan pregnancy and track pregnancy schedules until the estimated time of delivery. An example of an application that is widely used on the Google Play Store is FLO. FLO application that has been downloaded more than 100 million times and has been reviewed as many as 2 million reviews on the Google Play Store seen on January 12, 2023. Sentiment Analysis is an ongoing research field in the field of text mining and also the computational treatment of opinions, sentiments and subjectivity text that can be used as an evaluation of an application. The method chosen in this sentiment analysis research is CNN and RNN with Long Short Term Memory (LSTM) variants. In this study, the data used to carry out sentiment analysis is review data in text form. The results of sentiment analysis with the research object of FLO application reviews were 12,000 review data selected by country, namely Indonesia, which had more positive reviews, followed by neutral reviews and finally negative reviews. In the training data obtained, accuracy and loss by doing three epochs, namely 20, 50, 100 on the CNN and LSTM algorithms are good enough and not overfitting. Data testing is also carried out using confusion matrix and classification report, based on the two algorithm comparisons, the superior one is using the LSTM algorithm, with accuracy of 92.67% and 93% respectively, while the CNN accuracy results are 84.08% and 84% respectively.

Deep Convolutional Curvelet Transform- Based MRI Approach for Early Detection of ALZHEIMER' S Disease

Image processing is a technique for applying various procedures to a picture in order to improve it or extract some relevant information from it. It is a kind of signal processing where the input is an image and the output may either be another picture or features or characteristics related to that image. Image processing is one of the technologies that is currently expanding quickly. It is a primary subject of research in both the engineering and computer science fields. For dementia, Alzheimer's disease (AD) is the most prevalent kind. It often shows itself as a steadily declining memory and cognitive function, which makes it difficult for the affected person to live independently and has a significant negative influence on both the affected person and society. At the moment, AD diagnosis depends on medical history analysis, blood testing, behavior analysis, brain imaging, and cognitive testing. But because these processes are arbitrary and uneven, it is difficult to provide a precise forecast for the early phases of AD challenging. A dynamic dual-graph fusion convolutional network is suggested in this paper to enhance the accuracy of Alzheimer's disease (AD) diagnosis. The key contributions of the paper are as follows: The proposed architecture can dynamically adjust the graph structure for GCN to produce better diagnosis outcomes by learning the optimal underlying latent graph, propose a novel dynamic GCN architecture, which is an end-to-end pipeline for diagnosing AD, incorporate feature graph learning and dynamic graph learning, giving those useful features of subjects more weight while reducing the weights of other noise features. Experiments show that our approach achieves great classification results in AD diagnosis while offering flexibility and stability. This project is implemented for the Alzheimer disease classification using Graph convolutional network architecture like CNN with LSTM layer and fit the model in the training and testing and deploy the model for getting the test image output and then compare the accuracy of the model for getting the most perfect model for medical image classification and get the accuracy of CNN- LSTM layer with 98%.

A deep learning approach for Direct Immunofluorescence pattern recognition of Autoimmune Bullous Diseases.

Artificial intelligence (AI) is reshaping healthcare, using machine and deep learning to enhance disease management. Dermatology has seen improved diagnostics, particularly in skin cancer detection, through the integration of AI. However, the potential of AI in automating immunofluorescence imaging for autoimmune bullous skin diseases remains untapped. While direct immunofluorescence (DIF) supports diagnosis, its manual interpretation can hinder efficiency. The use of deep learning to automatically classify DIF patterns, including the Intercellular Pattern (ICP) and the Linear Pattern (LP), holds promise for improving the diagnosis of autoimmune bullous skin diseases. The objectives of this study are to develop AI algorithms for automated classification of autoimmune bullous skin disease DIF patterns, such as ICP and LP. This aims to enhance diagnostic accuracy, streamline disease management, and improve patient outcomes through deep learning-driven immunofluorescence interpretation. We collected immunofluorescence images from skin biopsies of patients suspected of AIBD between January 2022 and January 2024. Skin tissue was obtained via 5-mm punch biopsy, prepared for direct immunofluorescence. Experienced dermatologists classified the images into three classes: ICP, LP, and negative. To evaluate our deep learning approach, we divided the images into training (436) and test sets (93). We employed transfer learning with pre-trained deep neural networks and conducted 5-fold cross-validation to assess model performance. Our dataset's class imbalance was addressed using weighted loss and data augmentation strategies. The models were trained for 50 epochs using Pytorch, achieving an image size of 224x224 for both CNNs and the Swin Transformer. Our study compared six CNNs and the Swin transformer for AIBDs image classification, with the Swin transformer achieving the highest average validation accuracy of 98.5%. On a separate test set, the best model attained an accuracy of 94.6%, demonstrating 95.3% sensitivity and 97.5% specificity across AIBDs classes. Visualization with Grad-CAM highlighted the model's reliance on characteristic patterns for accurate classification. The study highlighted CNN's accuracy in identifying DIF features. This approach aids automated analysis and reporting, offering reproducibility, speed, data handling, and cost-efficiency. Integrating deep learning in skin immunofluorescence promises precise diagnostics and streamlined reporting in this branch of dermatology.

Comparison Accuracy of CNN and VGG16 in Forest Fire Identification: A Case Study

The current research aims to assess the precision of forest fire detection using CNN and VGG16 models, specifically in the context of fire identification. While both models have demonstrated significant promise in visual pattern recognition, a comprehensive analysis regarding their specific benefits in forest fire identification is still needed. The rationale behind this research stems from the significance of promptly identifying forest fires as a preemptive measure to mitigate their detrimental effects on the environment and society. The employed approach involves the application of transfer learning techniques on a diverse and extensive dataset encompassing different forest fire scenarios. The dataset was used to train both CNN and VGG16 models. The test results indicated that the CNN model achieved a forest fire detection accuracy of 96%, while VGG16 achieved 98% accuracy. The primary objective of this research is to enhance comprehension regarding the merits and demerits of each model in the context of forest fire identification scenarios. While VGG16 exhibits marginally superior performance in identifying forest fires, this discrepancy offers valuable insight into the practical applicability of these two models for fire detection in real-world scenarios. These findings establish a solid basis for the advancement of more dependable and efficient early detection technology in the prevention and management of forest fires in the future. This can be accomplished by capitalizing on the unique capabilities of each model to optimize their performance in practical scenarios.

Automated Staging of Diabetic Retinopathy Using Convolutional Support Vector Machine (CSVM) Based on Fundus Image Data

Diabetic Retinopathy (DR) is a complication of diabetes mellitus, which attacks the eyes and often leads to blindness. The number of DR patients is significantly increasing because some people with diabetes are not aware that they have been affected by complications due to chronic diabetes. Some patients complain that the diagnostic process takes a long time and is expensive. So, it is necessary to do early detection automatically using Computer-Aided Diagnosis (CAD). The DR classification process based on these several classes has several steps: preprocessing and classification. Preprocessing consists of resizing and augmenting data, while in the classification process, CSVM method is used. The CSVM method is a combination of CNN and SVM methods so that the feature extraction and classification processes become a single unit. In the CSVM process, the first stage is extracting convolutional features using the existing architecture on CNN. CSVM could overcome the shortcomings of CNN in terms of training time. CSVM succeeded in accelerating the learning process and did not reduce the accuracy of CNN's results in 2 class, 3 class, and 5 class experiments. The best result achieved was at 2 class classification using CSVM with data augmentation which had an accuracy of 98.76% with a time of 8 seconds. On the contrary, CNN with data augmentation only obtained an accuracy of 86.15% with a time of 810 minutes 14 seconds. It can be concluded that CSVM was faster than CNN, and the accuracy obtained was also better to classify DR.

The Implementation of the CNN Method on Smart Image Recognition and Identification of Heritage (SIRIH) of Sundanese Traditional Tools

Sundanese is a regional language in West Java. Young people rarely use local languages, so they need to be preserved by using technology. Image processing technology can help the process of strengthening the Sundanese language and culture. Researchers developed Sundanese traditional tools because their existence is not only related to Sundanese language and terms, but also related to philosophy and local wisdom values. This application is called Smart Image Recognition and Identification of Heritage (SIRIH). The development of this android-based innovative application is based on image recognition technology with the CNN method. Based on the test results, the classification accuracy of pre-trained CNN is 98%, and val accuracy at 86%. Accuracy test results on the application show good and maximum results (90%-100%) if the item is in sufficient light, both indoors and outdoors, while dim light will reduce accuracy to around 60%-70%. Blurred images also affect accuracy by up to 50%.

Automated Staging of Diabetic Retinopathy Using Convolutional Support Vector Machine (CSVM) Based on Fundus Image Data

Diabetic Retinopathy (DR) is a complication of diabetes mellitus, which attacks the eyes and often leads to blindness. The number of DR patients is significantly increasing because some people with diabetes are not aware that they have been affected by complications due to chronic diabetes. Some patients complain that the diagnostic process takes a long time and is expensive. So, it is necessary to do early detection automatically using Computer-Aided Diagnosis (CAD). The DR classification process based on these several classes has several steps: preprocessing and classification. Preprocessing consists of resizing and augmenting data, while in the classification process, CSVM method is used. The CSVM method is a combination of CNN and SVM methods so that the feature extraction and classification processes become a single unit. In the CSVM process, the first stage is extracting convolutional features using the existing architecture on CNN. CSVM could overcome the shortcomings of CNN in terms of training time. CSVM succeeded in accelerating the learning process and did not reduce the accuracy of CNN's results in 2 class, 3 class, and 5 class experiments. The best result achieved was at 2 class classification using CSVM with data augmentation which had an accuracy of 98.76% with a time of 8 seconds. On the contrary, CNN with data augmentation only obtained an accuracy of 86.15% with a time of 810 minutes 14 seconds. It can be concluded that CSVM was faster than CNN, and the accuracy obtained was also better to classify DR.

Fault Diagnosis of Rotating Machinery under Complex Conditions Based on Multi-Scale Convolutional Neural Networks

In complex operational scenarios involving variable speeds, burdens, and noise, rotating machinery necessitates extended maintenance. Extracting stable and effective fault-sensitive features under such intricate conditions presents a significant challenge. To tackle this issue, the paper introduces the Multi-Scale Convolutional Neural Network (MSCNN) model tailored specifically for such complexities. The approach in this paper simultaneously captures multi-scale vibration signal features using the innovative Multi-Scale Bifurcation (MSB) module and subsequently aggregates them through a multi-scale fusion layer. This model effectively addresses the common problem of low CNN accuracy in fault diagnosis. The paper validates methodology using a bearing dataset provided by Case Western Reserve University (CWRU) and demonstrates superior performance compared to classical CNN models and other alternatives, achieving an impressive 99.56% classification accuracy for normal signals.

DeepAutoGlioma: a deep learning autoencoder-based multi-omics data integration and classification tools for glioma subtyping

Background and objectiveThe classification of glioma subtypes is essential for precision therapy. Due to the heterogeneity of gliomas, the subtype-specific molecular pattern can be captured by integrating and analyzing high-throughput omics data from different genomic layers. The development of a deep-learning framework enables the integration of multi-omics data to classify the glioma subtypes to support the clinical diagnosis.ResultsTranscriptome and methylome data of glioma patients were preprocessed, and differentially expressed features from both datasets were identified. Subsequently, a Cox regression analysis determined genes and CpGs associated with survival. Gene set enrichment analysis was carried out to examine the biological significance of the features. Further, we identified CpG and gene pairs by mapping them in the promoter region of corresponding genes. The methylation and gene expression levels of these CpGs and genes were embedded in a lower-dimensional space with an autoencoder. Next, ANN and CNN were used to classify subtypes using the latent features from embedding space. CNN performs better than ANN for subtyping lower-grade gliomas (LGG) and glioblastoma multiforme (GBM). The subtyping accuracy of CNN was 98.03% (± 0.06) and 94.07% (± 0.01) in LGG and GBM, respectively. The precision of the models was 97.67% in LGG and 90.40% in GBM. The model sensitivity was 96.96% in LGG and 91.18% in GBM. Additionally, we observed the superior performance of CNN with external datasets. The genes and CpGs pairs used to develop the model showed better performance than the random CpGs-gene pairs, preprocessed data, and single omics data.ConclusionsThe current study showed that a novel feature selection and data integration strategy led to the development of DeepAutoGlioma, an effective framework for diagnosing glioma subtypes.

A Novel Approach for Improving the Security of IoT–Medical Data Systems Using an Enhanced Dynamic Bayesian Network

IoT (Internet of Things) devices are increasingly being used in healthcare to collect and transmit patient data, which can improve patient outcomes and reduce costs. However, this also creates new challenges for data security and privacy. Thus, the major demand for secure and efficient data-sharing solutions has prompted significant attention due to the increasing volume of shared sensor data. Leveraging a data-fusion-based paradigm within the realm of IoT-protected healthcare systems enabled the collection and analysis of patient data from diverse sources, encompassing medical devices, electronic health records (EHRs), and wearables. This innovative approach holds the potential to yield immediate benefits in terms of enhancing patient care, including more precise diagnoses and treatment plans. It empowers healthcare professionals to devise personalized treatment regimens by amalgamating data from multiple origins. Moreover, it has the capacity to alleviate financial burdens, elevate healthcare outcomes, and augment patient satisfaction. Furthermore, this concept extends to fortifying patient records against unauthorized access and potential misuse. In this study, we propose a novel approach for secure transmission of healthcare data, amalgamating the improved context-aware data-fusion method with an emotional-intelligence-inspired enhanced dynamic Bayesian network (EDBN). The findings indicated that F1 score, accuracy, precision, recall, and ROC-AUC score using DCNN were 89.3%, 87.4%, 91.4%, 92.1%, and 0.56, respectively, which was second-highest to the proposed method. On the other hand, the F1 score, accuracy, precision, recall, and ROC-AUC scores of FRCNN and CNN were low in accuracy at 83.2% and 84.3%, respectively. Our experimental investigation demonstrated superior performance compared with existing methods, as evidenced by various performance metrics, including recall, precision, F measures, and accuracy.

Lumpy Diseases Detection using Machine Learning

Machine learning algorithms for improving animal health monitoring have accelerated the creation of ML applications for behavioral and physiological monitoring systems, including ML-based animal health monitoring systems. Currently, farm animals are raised all over the world, and it is necessary to monitor their physiological processes. It is suggested in this article to use machine learning models to continuously monitor each animal's vital signs and look for biological changes. In this model, crucial data is gathered via IoT devices, and data analysis is carried out using machine learning techniques to identify potential dangers from changes in an animal's physiological state. The results of the experiments demonstrate that the suggested model is accurate and efficient enough to identify animal situations. For our purposes, the CNN and YOLO accuracy of more than 90% is a promising outcome. Keywords- Lumpy disease , Machine learning, Images

A Comparison of Machine Learning Algorithms for Music Genre Recognition

Music information retrieval includes the classification of music genres as a key component. However, due to the complexity of music composition, existing artificial classification cannot achieve accurate results. In the music genre classification scenario, machine learning can effectively process large and complex data, bringing more accurate and personalized results for music recommendation. This article will start from three articles and use GTZAN as a data set to compare 3 traditional machine learning methods, including SVM, random forest and logistic regression, together with 2 deep learning models, involving convolutional and feedforward neural networks (CNN and FFNN) in music genres. classification accuracy. The results show that the classification accuracy will be affected by multiple factors such as the characteristics of the audio, whether data processing is performed and so on. In addition, the performance of deep learning has not been significantly better than traditional machine learning as expected. At the same time, the accuracy of CNN and FFNN is heavily dependent on whether the audio is processed into a spectrogram.

An Enhancement Method in Few-Shot Scenarios for Intrusion Detection in Smart Home Environments

Different devices in the smart home environment are subject to different levels of attack. Devices with lower attack frequencies confront difficulties in collecting attack data, which restricts the ability to train intrusion detection models. Therefore, this paper presents a novel method called EM-FEDE (enhancement method based on feature enhancement and data enhancement) to generate adequate training data for expanding few-shot datasets. Training intrusion detection models with an expanded dataset can enhance detection performance. Firstly, the EM-FEDE method adaptively extends the features by analyzing the historical intrusion detection records of smart homes, achieving format alignment of device data. Secondly, the EM-FEDE method performs data cleaning operations to reduce noise and redundancy and uses a random sampling mechanism to ensure the diversity of the few-shot data obtained by sampling. Finally, the processed sampling data is used as the input to the CWGAN, and the loss between the generated and real data is calculated using the Wasserstein distance. Based on this loss, the CWGAN is adjusted. Finally, the generator outputs effectively generated data. According to the experimental findings, the accuracy of J48, Random Forest, Bagging, PART, KStar, KNN, MLP, and CNN has been enhanced by 21.9%, 6.2%, 19.4%, 9.2%, 6.3%, 7%, 3.4%, and 5.9%, respectively, when compared to the original dataset, along with the optimal generation sample ratio of each algorithm. The experimental findings demonstrate the effectiveness of the EM-FEDE approach in completing sparse data.

Audio-Visual CNN using Transfer Learning for TV Commercial Break Detection

The TV commercial detection problem is a hard challenge due to the variety of programs and TV channels. The usage of deep learning methods to solve this problem has shown good results. However, it takes a long time with many training epochs to get high accuracy. This research uses transfer learning techniques to reduce training time and limits the number of training epochs to 20. From video data, the audio feature is extracted with Mel-spectrogram representation, and the visual features are picked from a video frame. The datasets were gathered by recording programs from various TV channels in Indonesia. Pre-trained CNN models such as MobileNetV2, InceptionV3, and DenseNet169 are re-trained and are used to detect commercials at the shot level. We do post-processing to cluster the shots into segments of commercials and non-commercials. The best result is shown by Audio-Visual CNN using transfer learning with an accuracy of 93.26% with only 20 training epochs. It is faster and better than the CNN model without using transfer learning with an accuracy of 88.17% and 77 training epochs. The result by adding post-processing increases the accuracy of Audio-Visual CNN using transfer learning to 96.42%.

A study on the effect of different machine learning algorithms on soccer footwork recognition under trajectory tracking theory

Abstract This paper aims to investigate the effectiveness and influencing factors of different machine learning algorithms on soccer footwork recognition. In this paper, we use inertial sensors to obtain the basic data of soccer players’ movements, then convert them into initial data of footwork using pose representation and pose-solving filtering. The value of K mainly influences the classification accuracy of KNN, and the highest accuracy of 67.23% is achieved when K is 5. The classification accuracy of SVM is related to the choice of the distance function. The accuracy of CNN is mainly affected by the size of the convolutional kernel and the convolutional step size, and the highest accuracy is 73.82%. The machine learning-based soccer step recognition can improve the recognition accuracy of traditional physical methods and provide scientific sports guidance for soccer players’ training.

Retrieving Atmospheric Gas Profiles Using FY-3E/HIRAS-II Infrared Hyperspectral Data by Neural Network Approach

The observed radiation data from the second-generation Hyperspectral Infrared Atmospheric Sounder (HIRAS-II) on the Fengyun-3E (FY-3E) satellite contain useful vertical atmosphere information which can distinguish and retrieve vertical profiles of atmospheric gas components including ozone (O3), carbon monoxide (CO), and methane (CH4). This paper utilizes FY-3E/HIRAS-II observational data to optimize each gas channel using the improved Optimal Sensitivity Profile method (OSP) channel algorithm and establishes a typical convolutional neural network model (CNN) and a representative U-shaped network model (UNET) with deep features and shallow feature links to perform atmospheric profile retrieval calculations of O3, CO, and CH4. We chose the clear sky data of the Indian and its southern seas in December 2021 and January 2022, with reanalysis data from European Center for Medium-Range Weather Forecasts Reanalysis v5 (ERA5) and European Center for Medium-Range Weather Forecasts Atmospheric Composition Reanalysis v4 (EAC4) serving as the reference values. The retrieval outcomes were then compared against advanced numerical forecast models including the Whole Atmosphere Community Climate Model (WACCM), Global Forecast System (GFS), and satellite products from an Atmospheric Infrared Sounder (AIRS) and Infrared Atmospheric Sounding Interferometer (IASI). Experimental results show that the generalization ability and retrieval accuracy of CNN are slightly higher compared with UNET. For O3 profile retrieval, the mean percentage error (MPE) of the whole layers for CNN and UNET data in relation to ERA5 data was less than 8%, while the root-mean-square error (RMSE) was below 1.5 × 10−7 kg/kg; for CH4 profile retrieval, the MPE of the whole layers for CNN and UNET data in relation to EAC4 data was less than 0.7%, while the RMSE was below 1.5 × 10−8 kg/kg. The retrieval of O3 and CH4 are resulted in a significant improvement compared to the forecast data and satellite products in most pressure levels; for CO profile retrieval, the MPE of the whole layers for CNN and UNET data in relation to EAC4 data was less than 11%, while the RMSE was below 4 × 10−8 kg/kg. The error of the CO retrieval results was higher than that of the forecast data at the pressure level of 200~500 hPa and lower than that of similar satellite products with most pressure levels. The experiments indicated that the neural network method effectively determines the atmospheric gas profiles using infrared hyperspectral data, exhibiting a positive performance in accuracy and retrieval speed.

EEG-based analysis for pilots' at-risk cognitive competency identification using RF-CNN algorithm.

Cognitive competency is an essential complement to the existing ship pilot screening system that should be focused on. Situation awareness (SA), as the cognitive foundation of unsafe behaviors, is susceptible to influencing piloting performance. To address this issue, this paper develops an identification model based on random forest- convolutional neural network (RF-CNN) method for detecting at-risk cognitive competency (i.e., low SA level) using wearable EEG signal acquisition technology. In the poor visibility scene, the pilots' SA levels were correlated with EEG frequency metrics in frontal (F) and central (C) regions, including α/β (p = 0.071 < 0.1 in F and p = 0.042 < 0.05 in C), θ/(α + θ) (p = 0.048 < 0.05 in F and p = 0.026 < 0.05 in C) and (α + θ)/β (p = 0.046 < 0.05 in F and p = 0.012 < 0.05 in C), and then a total of 12 correlation features were obtained based on a 5 s sliding time window. Using the RF algorithm developed by principal component analysis (PCA) for further feature combination, these salient combinations are used as input sets to obtain the CNN algorithm with optimal parameters for identification. The comparative results of the proposed RF-CNN (accuracy is 84.8%) against individual RF (accuracy is 78.1%) and CNN (accuracy is 81.6%) methods demonstrate that the RF-CNN with feature optimization provides the best identification of at-risk cognitive competency (accuracy increases 6.7%). Overall, the results of this paper provide key technical support for the development of an adaptive evaluation system of pilots' cognitive competency based on intelligent technology, and lay the foundation and framework for monitoring the cognitive process and competency of ship piloting operation in China.

Design and Implementation of Convolutional Neural Network Accelerator Based on FPGA

Convolutional neural network, as a kind of feed-forward neural network, has been widely used in image recognition, speech processing and other fields in recent years. In this paper, an FPGA-based CNN gas pedal is designed to solve the problem of slow running and high-power consumption of CNN on resource-constrained hardware. The design invokes multi-stage pipeline parallel processing technology to accelerate convolutional operations; quantifies network parameters from 32-bit floating-point to 8-bit fixed-point while guaranteeing CNN accuracy, and uses data multiplexing to reduce resource consumption. Experimental results show that the design is 10 times faster than the intel i7-8700 and consumes only 1% of the power of the RTX 2060 at 50MHz.