阈值 Resources

Showing items tagged with "阈值"

A program implementing Particle Swarm Optimization (PSO) to enhance Backpropagation Neural Networks for classification tasks. The implementation follows a two-phase approach: first using PSO to optimize initial weights and thresholds, then training the BP network with momentum and adaptive learning rate algorithms. The attached materials include dataset and modular functions for data extraction, target generation, baseline BP implementation, PSO optimization, and integrated PSO-BP training.

MATLAB 191 views Tagged

ECG signal denoising based on wavelet transform with adjustable parameters including wavelet bases and thresholds to optimize experimental results. The implementation features detailed annotations, making it particularly valuable for beginners in this field. Code includes configurable wavelet families and thresholding methods for practical experimentation.

MATLAB 237 views Tagged

Traditional edge detection algorithms often struggle to establish appropriate thresholds for detecting faint edges with low gradient values in certain images. This implementation addresses this issue through two enhanced approaches: image variance normalization to amplify gradient values of blurred edges, and sigmoid function integration to incorporate regional pixel information into gradient calculations for threshold optimization. The program compares these advanced techniques with conventional edge detection methods, providing comprehensive performance evaluation through practical algorithm implementations using gradient computation and adaptive thresholding.

MATLAB 227 views Tagged

My improved adaptive segmentation approach incorporates the threshold obtained from Otsu's method as the initial threshold, effectively enhancing both processing speed and segmentation performance

MATLAB 189 views Tagged

In a MATLAB environment, a numerical simulation study was conducted using the db10 wavelet from the Daubechies wavelet family to compress speckle-degraded images. The research demonstrates that both threshold selection and decomposition levels significantly impact image compression quality. The implementation involves wavelet decomposition, thresholding operations, and reconstruction algorithms to analyze compression performance.

MATLAB 263 views Tagged

The wavelet threshold denoising method is a leading technique in image denoising that leverages the distinct characteristics of sub-band images after wavelet decomposition, applying different thresholds to achieve superior noise reduction results.

MATLAB 202 views Tagged