Laplacian Eigenmaps
Laplacian Eigenmaps is a manifold learning-based nonlinear dimensionality reduction technique that constructs weights using heat kernels, widely applicable in image segmentation to enhance clustering performance through spectral graph theory implementation.