A method for active learning of hyperspectral images (HSI) is proposed, which
combines deep learning with diffusion processes on graphs. A deep variational
autoencoder extracts smoothed, denoised features from a high-dimensional HSI,
which are then used to make labeling queries based on graph diffusion
processes. The proposed method combines the robust representations of deep
learning with the mathematical tractability of diffusion geometry, and leads to
strong performance on real HSI.