Premium
Extracting invariant characteristics of sketch maps: Towards place query‐by‐sketch
Author(s) -
Tang Ming,
Falomir Zoe,
Freksa Christian,
Sheng Yehua,
Lyu Haiyang
Publication year - 2020
Publication title -
transactions in gis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.721
H-Index - 63
eISSN - 1467-9671
pISSN - 1361-1682
DOI - 10.1111/tgis.12621
Subject(s) - sketch , invariant (physics) , consistency (knowledge bases) , computer science , matching (statistics) , perception , object (grammar) , artificial intelligence , information retrieval , geography , mathematics , algorithm , epistemology , statistics , philosophy , mathematical physics
In geography, invariant aspects of sketches are essential to study because they reflect the human perception of real‐world places. A person's perception of a place can be expressed in sketches. In this article, we quantitatively and qualitatively analyzed the characteristics of single objects and characteristics among objects in sketches and the real world to find reliable invariants that can be used to establish references/correspondences between sketch and world in a matching process. These characteristics include category, shape, name, and relative size of each object. Moreover, quantity and spatial relationships—such as topological, ordering, and location relationships—among all objects are also analyzed to assess consistency between sketched and actual places. The approach presented in this study extracts the reliable invariants for query‐by‐sketch and prioritizes their relevance for a sketch‐map matching process.