SpatialScene2Vec: A self-supervised contrastive representation learning method for spatial scene similarity evaluation

Open Webpage

Danhuai Guo, Yingxue Yu, Shiyin Ge, Song Gao, Gengchen Mai, Huixuan Chen

Published: 01 Jan 2024, Last Modified: 27 Sept 2024Int. J. Appl. Earth Obs. Geoinformation 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Achieving a cognitive understanding of spatial scenes at a comprehensive level.•Encoding of Spatial & Semantic Information with a holistic representation.•Contrastive learning is employed for spatial scene similarity Assessment.•Novel data augmentation method enhances the model’s robustness and variability.