Data Science
Hierarchical Image Segmentation - Home

Hierarchical Image Segmentation (HSEG) Home

The Hierarchical Image Segmentation (HSEG) software, created by Dr. James C. Tilton of NASA Goddard Space Flight Center, is designed to produce a set of hierarchically related segmentations from imagery data. These hierarchically related segmentations are at several levels of detail in which the coarser segmentations can be produced from the finer segmentations by selective merging of regions in the finer segmentations. A unique feature of this segmentation hierarchy is that all segmentation boundaries are delineated at the full spatial resolution of the imagery data (with no coarsening of image spatial resolution). This segmentation hierarchy can be used to identify objects, or regions of interest, in single band, multispectral or hyperspectral imagery data of up to 8,000 x 8,000 pixels in size (with parallel processing). Accompaning with the HSEG software is a viewing tool called HSEGViewer that allows the user to interact with the segmentation hierarchy produced by HSEG, including selecting which segmenation resolution is most appropriate, and labeling regions based on selecting specific data points. HSEG and HSEGViewer provide opportunites for improved remote-sensing (e.g., surveying of the environment and populations) and medical-imaging analysis (e. g. CAT scans, MRI, and ultrasound equipment).