GARE-Net: Geometric contextual aggregation and regional contextual enhancement network for image-text matching

Open Webpage

Fangming Zhong, Tao Zhou, Zhikui Chen, Suhua Zhang

Published: 01 Jan 2026, Last Modified: 05 Nov 2025Expert Syst. Appl. 2026EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•A novel GARE-Net consists of two key modules GCFA and RCFE for image-text matching.•Geometric Contextual Feature Aggregation (GCFA) tells where a given region is within the image.•Regional Contextual Feature Enhancement (RCFE) reflects what and where surrounding regions are.