CNN Feature Extractor Research Articles

Sustainable Construction of Higher Education MOOCs Using CNN Feature Extraction

The attention time of students studying in MOOC (Massive Open Online Courses) classroom was analyzed to optimize and further improve their performance. On this basis, a student class model based on convolutional neural networks (CNN) feature extraction was proposed. Through Pr (Adobe Premiere) technology, students' class videos were processed by framing, and relevant features were extracted based on changes in students' eye movement trajectories. Then, 10 class videos of ten different experimenters were selected for comparative experiments. After comparing the results, it was found that the test scores of the experimental personnel using MOOC model for assisted learning were significantly different from those before using MOOC model. The final test scores of the students using MOOC model for learning increased to 5-10 points, which had a certain positive impact on the learning results. In the context of sustainable development of higher education, the construction and application of the MOOC model require more favorable promotion and practice.

A nonlinear structural pulse-like seismic response prediction method based on pulse-like identification and decomposition learning

Abstract Accurate prediction of structural responses under seismic action is important for the performance evaluation of structures. However, depending on the differences in the characteristics of ground motions, particularly whether they contain impulses, the seismic-response characteristics of structures may differ significantly. Hence, a new method, i.e., WD–AttDL, is proposed in this study to predict structural response under pulse-like ground motions. This method innovatively combines a wavelet decomposition-based velocity pulse-identification method with decomposition learning, where decomposed pulses and high-frequency features are used as inputs to the neural-network model, thus simplifying the identification of pulse features for the model. The decomposition learning module integrates different types of submodules, such as CNN feature-extraction, LSTM temporal-learning, and self-attention mechanism submodules. To reduce the nonlinear time-range analysis calculations required to generate training data, a screening method for impulsive ground motions based on time-series feature clustering is proposed. The accuracy and applicability of the proposed method are verified by numerically simulating three different cycles of reinforced-concrete frame structures and a three-dimensional masonry structure. Compared with existing structural seismic-response models, WD–AttDL synergistically integrates different modules and thus offers a higher prediction accuracy. In particular, it reduces the peak error of the predicted response, which is important for the evaluation of structural performance. Additionally, based on the response predicted by WD–AttDL, a vulnerability analysis of the structure under pulse-like seismic effects can be performed rapidly and accurately.

Advanced thermal prediction for green roofs: CNN-LSTM model with SSA optimization

Green roofs serve as an innovative strategy for urban architecture, playing a pivotal role in mitigating urban heat island effects and enhancing building energy efficiency. This study proposes a hybrid model based on the Sparrow Search Algorithm (SSA) optimized Convolutional Neural Network − Long Short-Term Memory Network (CNN-LSTM) for rapid and accurate prediction of the thermal performance of green roofs. The model integrates the spatial feature extraction of CNN with the temporal analysis capability of LSTM, enhancing the predictive accuracy of multi-feature time series through global optimization by SSA. Taking the green roof in Guangzhou China as a case study and utilizing hourly meteorological and thermal property data, this model has demonstrated superior predictive performance: the Mean Absolute Error (MAE), Root Mean Square Error (RMSE), R-squared (R2), and Mean Absolute Percentage Error (MAPE) for the prediction of inner surface temperature are 0.718, 0.908 %, 0.922, and 0.025 %, respectively; for the outer surface temperature prediction, the corresponding values are 0.618, 0.886 %, 0.983, and 0.028 %. Comparative analysis with existing AI models and ablation experiments further confirm the superiority of this model. Additionally, through Spearman correlation analysis, this study reveals the correlation between key input parameters and surface temperatures, providing data support for the effective design and management of green roofs. The model presented in this study not only provides an efficient tool for predicting the thermal performance of green roofs but also holds significant importance for promoting green building and sustainable urban development.

Evaluation of precipitation forecasting methods and an advanced lightweight model

Abstract Precipitation forecasting is crucial for warning systems and disaster management. This study focuses on deep learning-based methods and categorizes them into three categories: Recurrent Neural Network (RNN-RNN-RNN), Convolutional Neural Network (CNN-CNN-CNN), and CNN-RNN-CNN methods. Then, we conduct a comprehensive evaluation of typical methods in these three categories using the SEVIR precipitation dataset. The results show that RNN-RNN-RNN suffers from instability in long-term forecasts due to error accumulation, CNN-CNN-CNN struggles to capture temporal signals but produces relatively stable forecasts, and CNN-RNN-CNN significantly increases model complexity and inherits the drawbacks of RNN, leading to worse forecasts. Here, we propose an advanced lightweight precipitation forecasting model (ALPF) based on CNN. Experimental results demonstrate that ALPF can effectively forecast spatial-temporal features, maintaining CNN’s feature extraction capabilities while avoiding error accumulation in RNN’s propagation. ALPF achieves long-term stable precipitation forecasts and can better capture large precipitation amounts.

CL-TransFER: Collaborative learning based transformer for facial expression recognition with masked reconstruction

Facial expression recognition (FER) has attracted intensive attention due to its critical role in various computer vision tasks. However, existing FER approaches suffer from either noisy annotations or expression ambiguity (high inter-class and low intra-class similarity), limiting the FER performance. To this end, we propose a robust end-to-end collaborative learning based transformer for FER (CL-TransFER) in this paper. Specifically, CL-TransFER co-trains a CNN feature extractor and a transformer feature extractor jointly to extract both rich local semantic features as well as global structural information from facial images. By enforcing the consensus between the predictions of two extractors, the CL-TransFER could suppress the influence of noisy annotations. To further tackle the expression ambiguity problem, we design a simple yet efficient self-supervised masked reconstruction (SSMR) task to pre-train the transformer feature extractor of CL-TransFER. This enhances the model's capability of learning fine-grained discriminative representations. Extensive experiments on three popular benchmarks have demonstrated the effectiveness and superiority of our method.

GM-DETR: Research on a Defect Detection Method Based on Improved DETR.

Defect detection is an indispensable part of the industrial intelligence process. The introduction of the DETR model marked the successful application of a transformer for defect detection, achieving true end-to-end detection. However, due to the complexity of defective backgrounds, low resolutions can lead to a lack of image detail control and slow convergence of the DETR model. To address these issues, we proposed a defect detection method based on an improved DETR model, called the GM-DETR. We optimized the DETR model by integrating GAM global attention with CNN feature extraction and matching features. This optimization process reduces the defect information diffusion and enhances the global feature interaction, improving the neural network's performance and ability to recognize target defects in complex backgrounds. Next, to filter out unnecessary model parameters, we proposed a layer pruning strategy to optimize the decoding layer, thereby reducing the model's parameter count. In addition, to address the issue of poor sensitivity of the original loss function to small differences in defect targets, we replaced the L1 loss in the original loss function with MSE loss to accelerate the network's convergence speed and improve the model's recognition accuracy. We conducted experiments on a dataset of road pothole defects to further validate the effectiveness of the GM-DETR model. The results demonstrate that the improved model exhibits better performance, with an increase in average precision of 4.9% (mAP@0.5), while reducing the parameter count by 12.9%.

Python Based Stress Detection System Using Machine Learning & Algorithm’s

In today's high-pressure world, the need for innovative, non-intrusive stress monitoring solutions is more critical than ever. This paper introduces a sophisticated stress detection system that combines facial expression analysis with machine learning to deliver real-time, continuous stress assessment. Our system employs Haar Cascade classifiers for precise face detection, CNN for feature extraction and utilizes the K-Nearest Neighbors (KNN) algorithm to classify stress-related facial expressions. A significant advantage of our system is its non-intrusive nature, relying solely on visual data, thus eliminating the need for cumbersome wearable sensors. This enhances user comfort and convenience. This paper outlines the design, implementation, and evaluation of our stress detection system, emphasizing its potential applications in workplace wellness, mental health monitoring, and human-computer interaction Key Words: - KNN, Facial Expressions, Machine Learning, Stress Detection, Image Processing, CNN

An Improved Index Price/Movement Prediction by using Ensemble CNN and DNN Deep Learning Technique

As it knows prediction is always a challenging task in all terms. The goal of any stock prediction method is to develop a robust method for predicting trading price that can be used to improve investment decisions and accurate models. The paper proposes a hybrid model that combines the strengths of deep learning models CNN and DNN, and to develop a comprehensive methodology for the prediction of stock/index prices on Banknifty (NSE Bank), a highly volatile Indian sectorial Index that represents 12 major banks of the country. The hybrid model consists of two main components a CNN for feature extraction and a DNN for regression or classification tasks. In the context of stock price prediction, CNN layers can be used to extract features from input data (such as stock prices and indicators) related to an estimated future value. DNN layer can be used to combine features learned from the CNN layers. Model performance will be evaluated using various metrics including Accuracy, Precision, Recall, Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared. The method also analyses the impact of various factors on stock prices, including market volatility, economic indicators, and geopolitical events. The achieved accuracy of 97.48% indicates that the model was successful in accurately predicting the stock prices of Bank Nifty. The proposed method is expected to provide investors and financial analysts with a valuable tool for making informed investment decisions.

Combining CNN Feature Extractors and Oversampling Safe Level SMOTE to Enhance Amniotic Fluid Ultrasound Image Classification

Combining CNN Feature Extractors and Oversampling Safe Level SMOTE to Enhance Amniotic Fluid Ultrasound Image Classification

QUIZ: An arbitrary volumetric point matching method for medical image registration

Rigid pre-registration involving local–global matching or other large deformation scenarios is crucial. Current popular methods rely on unsupervised learning based on grayscale similarity, but under circumstances where different poses lead to varying tissue structures, or where image quality is poor, these methods tend to exhibit instability and inaccuracies. In this study, we propose a novel method for medical image registration based on arbitrary voxel point of interest matching, called query point quizzer (QUIZ). QUIZ focuses on the correspondence between local–global matching points, specifically employing CNN for feature extraction and utilizing the Transformer architecture for global point matching queries, followed by applying average displacement for local image rigid transformation.We have validated this approach on a large deformation dataset of cervical cancer patients, with results indicating substantially smaller deviations compared to state-of-the-art methods. Remarkably, even for cross-modality subjects, it achieves results surpassing the current state-of-the-art.

Are you sure it’s an artifact? Artifact detection and uncertainty quantification in histological images

Modern cancer diagnostics involves extracting tissue specimens from suspicious areas and conducting histotechnical procedures to prepare a digitized glass slide, called Whole Slide Image (WSI), for further examination. These procedures frequently introduce different types of artifacts in the obtained WSI, and histological artifacts might influence Computational Pathology (CPATH) systems further down to a diagnostic pipeline if not excluded or handled. Deep Convolutional Neural Networks (DCNNs) have achieved promising results for the detection of some WSI artifacts, however, they do not incorporate uncertainty in their predictions. This paper proposes an uncertainty-aware Deep Kernel Learning (DKL) model to detect blurry areas and folded tissues, two types of artifacts that can appear in WSIs. The proposed probabilistic model combines a CNN feature extractor and a sparse Gaussian Processes (GPs) classifier, which improves the performance of current state-of-the-art artifact detection DCNNs and provides uncertainty estimates. We achieved 0.996 and 0.938 F1 scores for blur and folded tissue detection on unseen data, respectively. In extensive experiments, we validated the DKL model on unseen data from external independent cohorts with different staining and tissue types, where it outperformed DCNNs. Interestingly, the DKL model is more confident in the correct predictions and less in the wrong ones. The proposed DKL model can be integrated into the preprocessing pipeline of CPATH systems to provide reliable predictions and possibly serve as a quality control tool.

Gesture Recognition of sEMG Signals Based on CNN-GRU Network

To improve the accuracy of surface electromyogram signal (sEMG) gesture recognition algorithm and solve the problem of manually extracting many features, this paper proposes a deep neural network-based gesture recognition method. A neural network integrating CNN and GRU was designed. The 8-channel sEMG data collected by the MYO armband is input to the CNN for feature extraction, and then the obtained feature sequence is input to the GRU network for gesture classification, and finally the recognition result of the gesture category is output. The experimental findings that the proposed technology reaches 76.41% recognition accuracy on the MyoUP dataset. This demonstrates the practicality of the suggested plan.

Graph-Based Progressive Fusion Network for Multi-Modality Vehicle Re-Identification

Vehicle re-identification (Re-ID) is a critical task in intelligent transportation, aiming to match vehicle images of the same identity captured by non-overlapping cameras. However, it is difficult to achieve satisfactory results based on RGB images alone in darkness. Therefore, it is of great importance to consider multi-modality vehicle re-identification. Currently, the proposed works deal with different modality features through direct summation and fusion based on heat map, which however ignores the relationship between them. Meanwhile, there is a huge gap between the different modalities, which needs to be reduced. In this paper, to solve the above two problems, we propose a Graph-based Progressive Fusion Network (GPFNet) using a graph convolutional network to adaptively fuse multi-modality features in an end-to-end learning framework. GPFNet consists of a CNN feature extraction module (FEM), a GCN feature fusion module (FFM), and a loss function module (LFM). Firstly, in FEM, we employ a multi-stream network architecture to extract single-modality features and common-modality features and employ a random modality substitution module to extract mixed-modality features. Secondly, in FFM, we design an efficient graph structure to associate the features of different modalities and adopt a progressive two-stage strategy to fuse them. Finally, in LFM, we use GCN-aware multi-modality loss to constrain the features. For reducing modality differences and contributing better initial mixed-modality features to FFM, we propose random modality substitution as a data enhancement method for multi-modality datasets. Extensive experiments on multi-modality vehicle Re-ID datasets RGBN300 and RGBNT100 show that our model achieves state-of-the-art performance.

Multi-Method Analysis of Histopathological Image for Early Diagnosis of Oral Squamous Cell Carcinoma Using Deep Learning and Hybrid Techniques.

Oral cancer is a fatal disease and ranks seventh among the most common cancers throughout the whole globe. Oral cancer is a type of cancer that usually affects the head and neck. The current gold standard for diagnosis is histopathological investigation, however, the conventional approach is time-consuming and requires professional interpretation. Therefore, early diagnosis of Oral Squamous Cell Carcinoma (OSCC) is crucial for successful therapy, reducing the risk of mortality and morbidity, while improving the patient's chances of survival. Thus, we employed several artificial intelligence techniques to aid clinicians or physicians, thereby significantly reducing the workload of pathologists. This study aimed to develop hybrid methodologies based on fused features to generate better results for early diagnosis of OSCC. This study employed three different strategies, each using five distinct models. The first strategy is transfer learning using the Xception, Inceptionv3, InceptionResNetV2, NASNetLarge, and DenseNet201 models. The second strategy involves using a pre-trained art of CNN for feature extraction coupled with a Support Vector Machine (SVM) for classification. In particular, features were extracted using various pre-trained models, namely Xception, Inceptionv3, InceptionResNetV2, NASNetLarge, and DenseNet201, and were subsequently applied to the SVM algorithm to evaluate the classification accuracy. The final strategy employs a cutting-edge hybrid feature fusion technique, utilizing an art-of-CNN model to extract the deep features of the aforementioned models. These deep features underwent dimensionality reduction through principal component analysis (PCA). Subsequently, low-dimensionality features are combined with shape, color, and texture features extracted using a gray-level co-occurrence matrix (GLCM), Histogram of Oriented Gradient (HOG), and Local Binary Pattern (LBP) methods. Hybrid feature fusion was incorporated into the SVM to enhance the classification performance. The proposed system achieved promising results for rapid diagnosis of OSCC using histological images. The accuracy, precision, sensitivity, specificity, F-1 score, and area under the curve (AUC) of the support vector machine (SVM) algorithm based on the hybrid feature fusion of DenseNet201 with GLCM, HOG, and LBP features were 97.00%, 96.77%, 90.90%, 98.92%, 93.74%, and 96.80%, respectively.

Efficient Detection of Covid-19 from Chest X-ray Images using CNN Feature Extraction and an Ensemble of Machine Learning Classifiers

Efficient Detection of Covid-19 from Chest X-ray Images using CNN Feature Extraction and an Ensemble of Machine Learning Classifiers

A deep learning-based approach for distinguishing different stress levels of human brain using EEG and pulse rate

In today’s world, people suffer from many fatal maladies, and stress is one of them. Excessive stress can have deleterious effects on the health, brain, mind, and nervous system of humans. The goal of this paper is to design a deep learningbased human stress level measurement technique using electroencephalogram (EEG), and pulse rate. In this research, EEG signals and pulse rate of healthy subjects are recorded while they solve four different question sets of increasing complexity. It is assumed that the subjects undergo through four different stress levels, i.e., ‘no stress’, ‘low stress’, ‘medium stress’, and ‘high stress’, while solving these question sets. An attention mechanism-based CNN-TLSTM (convolutional neural network-tanh long short-term memory) model is proposed to detect the mental stress level of a person. An attention layer is incorporated into the designed TLSTM network to increase the classification accuracy of the CNN-TLSTM model. The CNN network is used for the automated extraction of intricate features from the EEG signals and pulse rate. Then TLSTM is used to classify the stress level of a person into four different categories using the CNNextracted features. The obtained average accuracy of the proposed CNN-TLSTM model is 97.86%. Experimentally, it is found that the designed stress level measurement technique is highly effective and outperforms most existing state-of-the-art techniques. In the future, functional Near-Infrared Spectroscopy (fNIRS), ECG, and Galvanic Skin Response (GSR) can be employed with EEG and pulse rate to increase the effectiveness of the designed stress level measurement technique.

White blood cell image analysis for infection detection based on virtual hexagonal trellis (VHT) by using deep learning

White blood cells (WBCs) are an indispensable constituent of the immune system. Efficient and accurate categorization of WBC is a critical task for disease diagnosis by medical experts. This categorization helps in the correct identification of medical problems. In this research work, WBC classes are categorized with the help of a transform learning model in combination with our proposed virtual hexagonal trellis (VHT) structure feature extraction method. The VHT feature extractor is a kernel-based filter model designed over a square lattice. In the first step, Graft Net CNN model is used to extract features of augmented data set images. Later, the VHT base feature extractor extracts useful features. The CNN-extracted features are passed to ant colony optimization (ACO) module for optimal features acquisition. Extracted features from the VHT base filter and ACO are serially merged to create a single feature vector. The merged features are passed to the support vector machine (SVM) variants for optimal classification. Our strategy yields 99.9% accuracy, which outperforms other existing methods.

VHXLA: A post-earthquake damage prediction method for high-speed railway track-bridge system using VMD and hybrid neural network

The timely and accurate prediction of post-earthquake damage is critically important for ensuring the safety of high-speed railway track-bridge systems. The study introduces a novel method known as the VHXLA model. This innovative model blends variational mode decomposition (VMD) with a hybrid neural network for predicting non-linear multi-component post-earthquake damage in high-speed railways. The VHXLA model comprises a signal processing module and a deep learning module. The signal processing module utilises VMD decomposition, and the Hilbert transform (HT) to transform two-dimensional seismic signals into four-dimensional complex time–frequency signals, thus simplifying the task of identifying seismic time–frequency characteristics via the neural network. The deep learning module integrates diverse types of neural network components, such as the Xception CNN feature extraction submodule, the LSTM RNN temporal learning submodule, and the multi-head attention mechanism submodule. A Bayesian self-optimisation method is implemented to determine the number of decomposition layers in VMD and select essential hyperparameters. The model's effectiveness is evaluated by predicting the damage of an experimentally validated finite element model, subjecting it to seismic loads, and thus gauging its performance. The results indicated that the signal processing module, based on VMD decomposition, significantly improves the neural network's signal processing capability. In addition, the synergistic integration of different modules in the VHXLA model provides superior prediction accuracy compared to pre-existing damage prediction models. Notably, the prediction accuracy is consistent across different positions of the same predicted component in the high-speed railway track-bridge system.

Semantic Knowledge Guided Class-Incremental Learning

Driven by practical needs, research on Class-Incremental Learning (CIL) has received more and more attentions in recent years. A technical challenge to be conquered by CIL methods is the catastrophic forgetting problem, where the model’s performance improves rapidly on new classes while deteriorates drastically on old ones. The main causes behind catastrophic forgetting include network drifts, inter-class confusions, etc. In this paper, we propose a novel CIL method that solves the catastrophic forgetting problem from two aspects. First, to solve the inter-class confusion problem, we propose a novel Semantic knOwledge gUided ciL framework (SOUL) that consists of a CNN feature extractor and a Bi-GCN (Graph Convolutional Network) classifier. In each CIL session, we use the semantic knowledge extracted from the class labels to build two inter-class relation graphs among all the encountered old and new classes. Using these two relation graphs, we develop a Bi-GCN classifier to fuse two kinds of semantic relations in a balanced way, and then to transfer the inter-class relations from semantic modality to image classification weights. The entire SOUL framework is trained end-to-end by the standard BP algorithm, which optimizes the Bi-GCN classifier and the CNN feature extractor jointly. Second, to prevent the network drift, we develop the local topology preserving strategy that divides the global topological structure of the learned feature space into a set of local topological relations, and maintains these local relations at CIL session. Experimental evaluations demonstrate the state-of-the-art performance accuracies on benchmark image classification datasets.

Adaptive Deep Feature Fusion for Continuous Authentication With Data Augmentation

Mobile devices are becoming increasingly popular and are playing significant roles in our daily lives. Insufficient security and weak protection mechanisms, however, cause serious privacy leakage of the unattended devices. To fully protect mobile device privacy, we propose ADFFDA, a novel mobile continuous authentication system using an Adaptive Deep Feature Fusion scheme for effective feature representation, and a transformer-based GAN for Data Augmentation, by leveraging smartphone built-in sensors of the accelerometer, gyroscope and magnetometer. Given the normalized sensor data, ADFFDA utilizes the transformer-based GAN consisting of a transformer-based generator and a CNN-based discriminator to augment the training data for CNN training. With the augmented data and the especially designed CNN based on the ghost module and ghost bottleneck, ADFFDA extracts deep features from the three sensors by the trained CNN, and exploits an adaptive-weighted concatenation method to adaptively fuse the CNN-extracted features. Based on the fused features, ADFFDA authenticates users by using the one-class SVM (OC-SVM) classifier. We evaluate the authentication performance of ADFFDA in terms of the efficiency of the transformer-based GAN, GAN-based data augmentation, CNN architecture, adaptive-weighted feature fusion, and OC-SVM classifier. The experimental results show that ADFFDA obtains the best authentication performance w.r.t representative approaches, by achieving a mean equal error rate of 0.01%.