Deep Diffusion Processes for Active Learning of Hyperspectral Images

01/08/2021 ∙ by Abiy Tasissa, et al. ∙ 0 ∙ share

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.

READ FULL TEXT

Spatially regularized active diffusion learning for high-dimensional images

An active learning algorithm for the classification of high-dimensional ...
11/06/2019 ∙ by James M. Murphy, et al. ∙ 12 ∙ share

read it
Active Deep Learning for Classification of Hyperspectral Images

Active deep learning classification of hyperspectral images is considere...
11/30/2016 ∙ by Peng Liu, et al. ∙ 0 ∙ share

read it
Learning by Active Nonlinear Diffusion

This article proposes an active learning method for high dimensional dat...
05/30/2019 ∙ by Mauro Maggioni, et al. ∙ 0 ∙ share

read it
Diffusion-based Deep Active Learning

The remarkable performance of deep neural networks depends on the availa...
03/23/2020 ∙ by Dan Kushnir, et al. ∙ 0 ∙ share

read it
Connecting exciton diffusion with surface roughness via deep learning

Exciton diffusion plays a vital role in the function of many organic sem...
10/31/2019 ∙ by Liyao Lyu, et al. ∙ 16 ∙ share

read it
Exploring the high dimensional geometry of HSI features

We explore feature space geometries induced by the 3-D Fourier scatterin...
03/01/2021 ∙ by Wojciech Czaja, et al. ∙ 0 ∙ share

read it
Spectral-Spatial Diffusion Geometry for Hyperspectral Image Clustering

An unsupervised learning algorithm to cluster hyperspectral image (HSI) ...
02/08/2019 ∙ by James M. Murphy, et al. ∙ 0 ∙ share

read it

Deep Diffusion Processes for Active Learning of Hyperspectral Images

Authors

Spatially regularized active diffusion learning for high-dimensional images

Active Deep Learning for Classification of Hyperspectral Images

Learning by Active Nonlinear Diffusion

Diffusion-based Deep Active Learning

Connecting exciton diffusion with surface roughness via deep learning

Exploring the high dimensional geometry of HSI features

Spectral-Spatial Diffusion Geometry for Hyperspectral Image Clustering

This week in AI