Feature Extraction

← Back to Notebook Examples

notebooks/feature_extraction/feature_extract_basics.ipynb

Pulls seven different feature maps out of three SEM images using spice.features.extract(). Covers: Gaussian smoothing at several scales, Sobel magnitude (edge strength), Gabor texture responses, Shannon entropy, the Frangi filter (highlights thin, vessel-like structures), and tensor coherence (how strongly aligned local texture is). Each spice.features.extract(feature_names=[...]) call builds a new SPICEData holding an (H, W, N_features) stack, sharing the source image’s history. The notebook extracts features one at a time in a loop, building up a shared history, then shows each feature map with a title describing it.