Image Segmentation Using Bimodal Method with Histogram Display
Image segmentation using bimodal method with histogram display, determining threshold based on histogram analysis, and applying this threshold for image stylization processing
Estimation of Threshold Values at Various Scales
In partial discharge testing, acquired signals often contain white noise and periodic interference that need removal. This implementation utilizes the commonly used db6 wavelet from the Daubechies series to perform a 9-level multiresolution decomposition. Based on the energy characteristics of white noise, threshold values for each scale are estimated using hard thresholding processing, followed by signal reconstruction. The algorithm involves wavelet decomposition, noise variance estimation, and threshold application using MATLAB's wdenoise function or custom implementation with wthresh.
Genetic Algorithm Optimized BP Neural Network Weight and Threshold Parameters
This MATLAB program implements genetic algorithm optimization for BP neural network weight and threshold parameters, featuring population initialization, fitness evaluation, crossover, mutation operations, and neural network training integration.
Image Segmentation Techniques
Implementation of threshold-based image segmentation methods using MATLAB, including Otsu's method, maximum entropy, clustering approaches, and other algorithms with practical usage examples
Iterative Optimal Threshold Segmentation Algorithm
Iterative optimal threshold segmentation algorithm for edge detection by finding the best threshold value, with implementation-focused explanations of key computational steps and evaluation metrics.
Speech Denoising Based on Wavelet Packet Decomposition
Speech denoising using wavelet packet decomposition with threshold estimation via entropy spectral probability density function to remove real-world environmental noise. The implementation involves signal decomposition, entropy-based threshold calculation, and coefficient processing for noise removal.
Threshold-Based Particle Comparison Criterion for Multi-Objective Constrained Optimization
A threshold-based particle comparison criterion for handling multi-objective constrained optimization problems, which retains infeasible solution particles with small ordinal values and constraint violations within acceptable ranges, facilitating evolution from infeasible to feasible solutions. A novel crowding distance function assigns higher values to points in sparse regions and near Pareto front boundaries, increasing their selection probability, constituting a hybrid particle swarm optimization algorithm for solving multi-objective constrained optimization problems.
Threshold Selection in Wavelet Transform
Wavelet transform-based threshold selection for noise reduction, demonstrating the superior performance of soft thresholding over hard thresholding in signal denoising applications.
Characteristics of Wavelet Transform Multiscale Analysis for Enhanced Speech Denoising
Leveraging the multiscale analysis properties of wavelet transform, this approach improves basic frequency-domain spectral subtraction by applying different thresholds at various wavelet domain scales. It separates voiced and unvoiced sounds during denoising based on their distinct characteristics, preserving unvoiced components to produce fuller-sounding speech while employing adaptive algorithms to further enhance signal quality.
Retinal Blood Vessel Extraction Using Kirsch's Edge Detection Templates
This program implements blood vessel extraction from retinal images using Kirsch's edge detection templates. The algorithm applies Kirsch's templates at multiple orientations to filter input retinal images, with an adjustable threshold parameter for fine-tuning the vessel extraction output. The implementation includes directional filtering and morphological processing to enhance vascular structure detection.