Fig. 1: Two different frameworks to identify the correlation between polarization and pathological features. | Communications Engineering

Fig. 1: Two different frameworks to identify the correlation between polarization and pathological features.

From: Unsupervised learning of pixel clustering in Mueller matrix images for mapping microstructural features in pathological tissues

Fig. 1

The top path corresponds to the supervised learning approach, where polarization features are extracted for the pathological labels. The bottom path corresponds to the unsupervised learning approach, where the polarization pixels are first clustered into subtypes, and the subtypes that correlate with pathological variation are identified.

Back to article page