以高頻譜影像分析為例
Hyperspectral Image Analysis
In the linear mixture model for hyperspectral images, all the image spectra lie on a high-dimensional simplex with corners called endmembers. Given a set of endmembers, the standard calculation of fractional abundances with constrained least squares typically identifies the spectra as combinations of most, if not all, endmembers. We instead assume that pixels are combinations of only a few endmembers, yielding abundance vectors that are sparse. We introduce Sparse Demixing (SD), a method, similar to Orthogonal Matching Pursuit (OMP), for calculating these sparse abundances. We demonstrate that SD outperforms an existing L1 demixing algorithm, which we prove depends adversely on the angles between endmembers. We combine SD with dictionary learning methods to automatically calculate endmembers for a provided set of spectra. Applying to an AVIRIS image of Cuprite, Nevada yields endmembers that compare favorably with signatures from the USGS spectral library.
Keywords: hyperspectral; demix; hsy; hyperspectral image; image; sparse; omp; abundance; spectra; usg; match pursuit; spectral library; dictionary; mixture model; nevada;
Tongue Tumor Detection in Medical Hyperspectral Images
Extreme Learning Machines (ELM)
網易公開課程
Exploiting Sparsity in Hyperspectral Image Classification via Graphical Models
Abstract—A significant recent advance in hyperspectral image
classification relies on the observation that the spectral signature
of a pixel can be represented by a sparse linear combination of
training spectra which come from an over-complete dictionary.
以機率密度函數估計為例
NONPARAMETRIC ESTIMATION OF A MULTIVARIATE PROBABILITY DENSITY FOR MIXING SEQUENCES BY THE METHOD OF WAVELETS
Density estimation by wavelet thresholding
Some New Methods for Wavelet Density Estimation
wavelet density estimation
mixture model density estimation
Density Estimation and Mixture Models
gaussian mixture model density estimation
Expectation-Maximization (EM) algorithmGaussian mixture model (Mixture of Gaussians)
Estimating Gaussian Mixture Densities with EM – A Tutorial
Model selection - introduction and examplesLecture 08, part 1 | Pattern Recognition
the difficulty is then to select the appropriate Hilbert space E. And this is where and why Sobolev
spaces appeared.
Chapter 1: Sobolev Spaces Introduction
L. Rudin, S. Osher, and E. Fatemi. Nonlinear Total Variation based
Noise Removal Algorithms. Physica D, 60:259-268, 1992.
Recent Developments in Total Variation Image
Restoration
Lecture Notes on Bayesian Estimation and Classification
An Introduction to Conditional Random
Fields for Relational Learning
Restoration of Degraded Images with Maximum Entropy
Conditional random fields: Probabilistic
models for segmenting and labeling sequence data
Markov Random Fields
