Vogel divide the image into 10*10 grids resulting in 100 patches.
These patches are then annotated using nine local concepts.
He build a model that categorize images into six major categories namely forest, coasts, sky_clouds, water scapes, landscape with mountains, field.
Zia worked for queries like find images with mountains and sky on top.
He investigated three types of spatial relationships that is relative size of each of the concept occurrences, Allen’s calculus for describing the 13 relations namely before, meet, contains etc between the concepts.
Also he had applied chord pattern to each grid row.