Python notebooks

The following python notebooks contain examples demonstrating Higra usage. Data can be points, images, or meshes, or anything that can be transformed into a graph.

Component trees

These examples deal with upper and lower treshold set of vertex-weighted graphs.

Connected image filtering with component trees	👁	💾
Filtering with non-increasing criterion - The shaping framework	👁	💾
Pattern spectra - granulometry based on connected filters	👁	💾

Hierarchical segmentation

These examples deal with images, weights are on edges of the associated graph.

Visualizing hierarchical image segmentation	👁	💾
Watershed hierarchies	👁	💾
Hierarchy filtering	👁	💾
Computing a saliency map with the shaping framework	👁	💾
Multiscale Hierarchy Alignment and Combination	👁	💾

Triangular meshes

We provide two examples.

The first one uses trimesh, a simple, pure-python. It can be slow, and not-memory efficient.
The second one uses igl, an efficient C++ geometry processing library, with python bindings.

Hierarchical mesh segmentation – trimesh	👁	💾
Hierarchical mesh segmentation – igl	👁	💾

Useful tools

Region Adjacency Graph	👁	💾
Interactive object segmentation	👁	💾
Contour Simplification	👁	💾

Illustrative applications from scientific papers

Points and Images - Illustrations of SoftwareX 2019 article	👁	💾
Non-relevant node removal, on both point and images. PRL 2019	👁	💾
Fuzzy-marker-based interactive object segmentation - DGMM 2021	👁	💾
Astronomical object detection with the Max-Tree - MMTA 2016	👁	💾