Beyond the geotag: assessing implicit geoprivacy risks in visual user-generated content

doi:10.1080/13658816.2026.2623524

CSpace

	Beyond the geotag: assessing implicit geoprivacy risks in visual user-generated content
	Cheng, Shengjia 1; Chen, Peng 1; Zhu, Longsheng 1; Yan, Yu 2
	2026-02-04
发表期刊	INTERNATIONAL JOURNAL OF GEOGRAPHICAL INFORMATION SCIENCE
ISSN	1365-8816
页码	26
摘要	User-Generated Content (UGC) on social media has created a wealth of geographic information. However, it also presents a significant geoprivacy challenge. Image geolocation is the task of identifying an image's location. It offers benefits in domains such as disaster response and autonomous driving. Concurrently, it enables the inference of personal locations from implicit geographic cues within the images. Thus, quantitatively assessing the implicit geoprivacy risk of images is becoming increasingly important. To address this problem, this study constructs a Bayesian Network that includes (1) a comprehensive feature set of 11 metrics in four categories of an image's geolocation, (2) a taxonomy of five core geolocation methods, modeling the causal pathways from the features to the methods, and finally to the vulnerability. We constructed a dataset with 42,576 real-world examples of image geolocation. The model's AUC reaches 0.90 on the test dataset (where 1.0 is perfect and 0.5 is random chance). The results also support that task attractiveness, scene complexity, and landmark level are vital factors in determining the final localization probability. This study provides granular insights into how specific features influence the choice of localization pathways.
关键词	Geoprivacy geo-location risk assessment bayesian network volunteered geographic information
DOI	10.1080/13658816.2026.2623524
收录类别	SCI
语种	英语
WOS研究方向	Computer Science ; Geography ; Physical Geography ; Information Science & Library Science
WOS类目	Computer Science, Information Systems ; Geography ; Geography, Physical ; Information Science & Library Science
WOS记录号	WOS:001681541600001
出版者	TAYLOR & FRANCIS LTD
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/42812
专题	中国科学院计算技术研究所
通讯作者	Chen, Peng
作者单位	1.Peoples Publ Secur Univ China, Sch Informat & Cyber Secur, Beijing, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Cheng, Shengjia,Chen, Peng,Zhu, Longsheng,et al. Beyond the geotag: assessing implicit geoprivacy risks in visual user-generated content[J]. INTERNATIONAL JOURNAL OF GEOGRAPHICAL INFORMATION SCIENCE,2026:26.
APA	Cheng, Shengjia,Chen, Peng,Zhu, Longsheng,&Yan, Yu.(2026).Beyond the geotag: assessing implicit geoprivacy risks in visual user-generated content.INTERNATIONAL JOURNAL OF GEOGRAPHICAL INFORMATION SCIENCE,26.
MLA	Cheng, Shengjia,et al."Beyond the geotag: assessing implicit geoprivacy risks in visual user-generated content".INTERNATIONAL JOURNAL OF GEOGRAPHICAL INFORMATION SCIENCE (2026):26.