Feature Extraction and Data Modeling of Multi-Frequency Electronic Tongue Signals for Monitoring the Processing Stages of Ginger-Processed Pinellia ternata (Zhejiang)

International audience

Automatic identification of flaws is very important for ultrasonic nondestructive testing and evaluation of pipelines.A novel automatic identification approach of flaws using support vector machine (SVM) is presented.Wavelet transform is applied to feature extraction of ultrasonic echo signals, and SVM is to perform the identification task.To validate this approach, some experiments are performed.The results show that unlike conventional and artificial neural networks (ANN) identification methods the new technique performs better than conventional evaluation ones with advantages of high identification performance for pipeline flaws, lower cost, excellent generalization.

In order to improve the accuracy of EEG signals extraction and classification, a new feature extraction and classification method of EEG signals based on multi-task motor imagination is proposed. First, the empirical mode decomposition method is used to decompose the EEG signal into several intrinsic mode functions, and then the fast independent component analysis algorithm is used for the obtained intrinsic mode functions, so as to finally obtain the denoised EEG signal. Aiming at the four types of task modes of motor imagery in brain-computer interaction, this paper adopts the feature extraction algorithm of wavelet packet decomposition (WPD) fusion common spatial pattern (CSP). First, the input EEG signal is wavelet packet decomposition, and then the CSP feature extraction method is used to perform feature extraction for the four types of EEG signals under the “one-to-many” strategy. In terms of classification, a four-task classification method based on support vector machine (SVM) is designed. The experimental results show that the classification accuracy of the four types of motor imagery EEG signals using this paper method is 78.3%, compared with the pure CSP feature extraction, the accuracy rate is increased by 7.8%. It proves that the denoising method, feature extraction and classification method proposed in this paper are effective.

This study aims to compare the results of four feature extraction models in the case of early recognition of disease attacks on cocoa fruits. The image extraction models used in this study are Local Binary Pattern (LBP), Gray Level Co-occurrence Matrix (GLCM), Hue Saturation Value (HSV), and Gray-level Co-occurrence Histograms (GLCH). In addition, the Support Vector Machine (SVM) model was used for the classification technique to evaluate the extraction results from the cocoa image dataset. The classification results using SVM showed the best performance on feature extraction HSV in all types of Kernel SVM used (Linear, RBF, and Polynomial), with the highest accuracy of 80.95% on RBF Kernel. Furthermore, the HSV performance in recognizing disease attacks on cocoa fruits, based on Precision, Recall, and F1-Score values, showed that, on average, HSV had a better value than other feature extraction methods.

As an innovative communication method, brain-computer interface (BCI) can directly convert human brain activity into control signals, which is of great significance in improving the quality of life of people with disabilities. In this paper, the application of steady-state visual evoked potential (SSVEP) in BCI is discussed, and the feature extraction and classification methods of EEG signals are studied. Feature extraction of EEG signals was performed by preprocessing them using the EEGLAB toolbox and classification using support vector machine (SVM) to identify different patterns of EEG activity. The experimental results show that this feature extraction and classification method significantly improves the performance of BCI system. Future research can further optimize the feature extraction algorithm and improve the visual stimulation paradigm to improve the recognition accuracy and practicality of the system. Additionally, integrating advanced machine learning techniques such as deep learning and transfer learning could potentially enhance the system’s ability to adapt to individual users and generalize across different tasks, thereby increasing the robustness and versatility of the BCI system in real-world applications.

Electroencephalogram (EEG) is a non-stationary random signal with strong background noise, which makes its feature extraction difficult and recognition rate low. This paper presents a feature extraction and classification model of motor imagery EEG signals based on wavelet threshold denoising. Firstly, this paper uses the improved wavelet threshold algorithm to obtain the denoised EEG signal, divides all EEG channel data into multiple partially overlapping frequency bands, and uses the common spatial pattern (CSP) method to construct multiple spatial filters to extract the characteristics of EEG signals. Secondly, EEG signal classification and recognition are realized by the support vector machine algorithm optimized by a genetic algorithm. Finally, the dataset of the third brain-computer interface (BCI) competition and the dataset of the fourth BCI competition is selected to verify the classification effect of the algorithm. The highest accuracy of this method for two BCI competition datasets is 92.86% and 87.16%, respectively, which is obviously superior to the traditional algorithm model. The accuracy of EEG feature classification is improved. It shows that an overlapping sub-band filter banks common spatial pattern-genetic algorithms optimization-support vector machines (OSFBCSP-GAO-SVM) model is an effective model for feature extraction and classification of motor imagination EEG signals.

At present, the analysis of UAV flight acoustic signals is mainly based on traditional speech signal processing methods, and has not been analyzed in depth. According to the flight signal of UAV, combined with the aerodynamic characteristics of UAV, the characteristics of UAV’s acoustic signal are analyzed. The three feature extraction algorithms of pitch period, FFT and Mel Cepstral Coefficient (MFCC) are analyzed and compared. Feature extraction is performed, and a support vector machine (SVM) classification algorithm is applied to perform multi-classification model recognition. The measured and experimental results show that based on SVM classification and recognition, the three feature recognition methods all realize the classification of the model. The comprehensive FFT is the best, and the MFCC is the second. The pitch period is not suitable as the feature extraction method alone.

Nowadays surface electromyography (sEMG) signals play a very authoritative role in facilitating a neuromuscular disordered person and disabled person to live a smooth life. This offline study mainly focuses on the denoising, feature extraction and classification of the sEMG signals with discrete wavelet packet transform (DWT) with ensemble support vector machine (SVM) classification approach. In this work DWT (db 2, 4th level) is selected for denoising and TFD feature extraction to form feature vectors and soft thresholding method (minimax) was utilized and the threshold value was taken as 2.991. The classification accuracy is 98% is achieved at the better precision and speed of response for elbow movement. feature vectors and soft thresholding method (minimax) was utilized and the threshold value was taken as 2.991. The classification accuracy is 98% is achieved at the better precision and speed of response for elbow movement.

In modern times, the movie industry is growing rapidly. Netflix is one of the platforms that can be used to watch movies and provides many types of genres and movie titles. With so many genres and movie titles sometimes making it difficult for people to choose a movie to watch, one solution to the problem is a recommendation system that can recommend movies based on user ratings. One method in the recommendation system is collaborative filtering. One of the algorithms contained in collaborative filtering is singular value decomposition. Twitter is one of the places where people often write their opinions about the movies they have watched, from people's tweets on Twitter will be processed into rating value data. In this system, tweets become input that is processed into data that has a rating. This research implements a user-based recommendation system based on ratings from tweets using collaborative filtering combined with the Singular Value Decomposition (SVD) algorithm and Support Vector Machine (SVM) classification and implemented it on user-based and item-based. This research aims to implement a system that combines collaborative filtering techniques with the Singular Value Decomposition (SVD) algorithm and Support Vector Machine (SVM) classification. With the hope of producing a good movie recommendation model and providing accurate predictions for recommended and non-recommended movies. The test results in this study show that Collaborative Filtering gets the best RMSE value of 0.8162 on user-based and 0.5911 on item-based. The combination of Singular Value Decomposition (SVD) algorithm and Support Vector Machine (SVM) classification using hyperparameter tuning resulted in 81% precision and 81% recall for user-based while 80% precision and 80% recall for item-based.

The present work was aimed to present a thorough experimental study that shows the superiority of the generalisation capability of the support vector machine (SVM) approach in the classification of electrocardiogram (ECG) signals. Feature extraction was done using principal component analysis (PCA). Further, a novel classification system based on particle swarm optimisation (PSO) was used to improve the generalisation performance of the SVM classifier. For this purpose, we have optimised the SVM classifier design by searching for the best value of the parameters that tune its discriminant function and upstream by looking for the best subset of features that feed the classifier. The obtained results clearly confirm the superiority of the SVM approach as compared to traditional classifiers, and suggest that further substantial improvements in terms of classification accuracy can be achieved by the proposed PSO-SVM classification system.

The real estate market has a very important role in our society. It has a relationship with development and a person’s fundamental need. So, correct forecasting for sales and demand for real estate is very significant. SVM is a generous type of learning machine which solves classification with limited sample learning, nonlinear classification as well as handle “curse of dimensionality”. SVM has powerful classification capability with the feature selection, kernel selection, and parameter optimization add-on the classification accuracy. This paper focus on real estate sales forecasting and booking scenario on the basis of customer enquiry features. Paper follows the approach of Support Vector Machine (SVM) classification to forecast sales in real state. SVM is a type of machine learning algorithm from this, inference knowledge for prediction of sale. The proposed model helps real estate people to make a decision for the further stage of the construction or launch a new project according to sales and demand. For the classification, data is gathered from the real estate project. SVM classification accuracy is measured with polynomial kernel and feature selection. The optimal solution can be found and forecasting effect can be achieved by SVM classification. The experimental result proves that the SVM has good forecasting capability. Results also identify that how classification in real estate provides the solution for sales forecasting.KeywordsSales forecastingReal estateKernelFeatureSupport vector

Heart sound is an important basis for analyzing heart health. The feature extraction and classification model of heart sound signal are optimized, and an auxiliary diagnosis algorithm for congenital heart disease based on multi-modality and dense residual network is proposed. This method does not need to segment heart sounds, but only needs to extract double features in the time domain, which simplifies the process of preprocessing and feature extraction. Double features are learned using a parallel double-branch convolutional recurrent neural network with dense residual connections. The proposed classification algorithm was trained, validated, and tested on a total of 4050 clinical congenital heart disease heart sound samples, and obtained a classification accuracy of 97.60%, which is expected to be used for clinical auxiliary diagnosis and screening of congenital heart disease.

In this paper, a Histogram of Oriented Gradients (HOG) feature extraction accelerator for real-time multiple object detection is presented. The processor employs three techniques: a VLSI-oriented HOG algorithm with early classification in Support Vector Machine (SVM) classification, a dual core architecture for parallel feature extraction, and a detection-window-size scalable architecture with a reconfigurable MAC array for processing objects of different shapes. Early classification reduces the number of computations in SVM classification. The dual core architecture and the detection-window-size scalable architecture enable the processor to operate in several modes: high-speed mode, low-power mode, multiple object detection mode, and multiple shape object detection mode. These techniques expand the processor flexibility required for versatile application. The test chip was fabricated using 65 nm CMOS technology. The proposed architecture is designed to process HDTV resolution video (1920 × 1080 pixels) at 30 frames per second (fps). The performance of this accelerator is demonstrated on a pedestrian detection system.

Classification and identification of yellowing, spotted and normal green leaves were performed by using hyperspectral imaging technology. Firstly, three kinds of support vector machine (SVM) classification models were established under different preprocessing methods. The optimal preprocessing method (normalized+ SG+ first derivative) was obtained by using the pixel as the data set input and the recognition rate of the classification reached 97.4%. However, the amount of data information is large and the detection time is relatively long. In order to further optimize the spectral characteristics and improve the detection efficiency, principal component analysis (PCA) was used for feature selection and feature extraction under this optimal pretreatment method. The experiment shows that the feature extraction method selects the optimal principal component number of 7 in the dimension reduction experiment with the dimension ranging from 3 to 15 and the recognition rate is 97.3% after SVM classification, which is 4.8 percentage points higher than the feature selection method. The method of feature selection speeds up by 22.57s, which is faster than the optimal preprocessing method by 104.07s, which can greatly improve the recognition and detection speed while ensuring the accuracy. The results show that the hyperspectral image technology combined with the support vector machine classification model can achieve the non-destructive rapid detection of green leaf defects under the normalization + SG + first derivative preprocessing method and feature extraction.

Detection and classification of subject-generated artifacts in EEG signals using autoregressive models