Fig. 11 | Scientific Reports

Fig. 11

From: Enhancing feature learning of hyperspectral imaging using shallow autoencoder by adding parallel paths encoding

Fig. 11

t-SNE visualization of the balanced training set before and after feature encoding: The visualizations illustrate that the distribution and clustering patterns of the original and reduced features are nearly identical, indicating that the AE effectively preserves the essential data structure while reducing dimensionality. This reflects the AE’s ability to retain meaningful information and remove redundancy without compromising the class-separating characteristics of the data.

Back to article page