特征提取 Resources

Showing items tagged with "特征提取"

Early detection of arrhythmia is critical for cardiac patients through electrocardiogram (ECG) signal analysis and feature extraction. This study implements three distinct feature extraction algorithms—Fast Fourier Transform (FFT), Autoregressive (AR) Modeling, and Principal Component Analysis (PCA)—combined with an Artificial Neural Network (ANN) classifier. The PCA-based system achieved superior accuracy of 92.7083% using 3-second ECG intervals, outperforming reference methods (84.4%). The approach demonstrates scalable applicability for arrhythmia classification and potential cardiac disease prediction.

MATLAB 207 views Tagged
Download

Development and enhancement of multifractal spectrum algorithms and box-counting dimension calculation methods using MATLAB software, applicable for mechanical equipment fault diagnosis and feature extraction. These implementations provide valuable insights for applying fractal theory to diagnostic systems, featuring optimized code structure with efficient matrix operations and custom functions for partition-based analysis.

MATLAB 217 views Tagged
Download

KPCA and SVM combined for face recognition - SVM enhances classification performance while KPCA provides superior feature extraction using kernel functions inspired by SVM methodology

MATLAB 216 views Tagged
Download

A MATLAB-based handwritten digit recognition system utilizing Backpropagation Neural Network, featuring interactive input canvas, feature extraction algorithms, neural network model training, and real-time digit classification capabilities. Complete implementation details and usage instructions are provided in the README documentation.

MATLAB 256 views Tagged
Download