Call for papersÂ
Submission CMT*: https://cmt3.research.microsoft.com/CV4EO2025/
Author guidelines (follow WACV main conference): https://wacv2025.thecvf.com/submissions/author-guidelines/
Deadlines:
- Paper submission deadline: 30 November, 2024 (23:59, Pacific Time)
- Notification deadline: 30 December, 2024
- Camera-ready papers: 10 January 2025
- Workshop: 28 February 2025 (in person with limited virtual participation)
Click to Submit
We welcome submissions of full papers as well as position papers, work-in-progress, and papers discussing open problems and challenges. Topics include, but are not limited to:
Computer vision models for geospatial data
- Multimodal geospatial data integration/fusion, ranging from different image resolutions, viewpoints (e.g., off-nadir imagery or ground-to-aerial matching), sensors, modalities, as well as combination of raster and vector data. We particularly encourage submissions addressing geospatial-awareness and modalities other than RGB optical imagery;
- Architectures, loss functions, contrastive learning and other mechanisms leveraging spatiotemporal components and/or metadata available from geospatial data;
- Self-/semi-supervised learning schemes toward unimodal and multimodal Foundation Models (Large Vision – Language Models) for Earth observation applications.
Applications and workflows
- Novel applications of computer-vision techniques for geospatial imagery analysis, e.g. humanitarian assistance, precision agriculture, sustainable development goals, national security missions, environment/wildlife monitoring;
- Methods for harmonization, homogenization, and tokenization of varied remote sensing data types (e.g., multispectral, hyperspectral, SAR data) toward Analysis Ready Data (ARD) and AI-ready data;
- Metrics and mechanisms for data/model visualization, uncertainty quantification, improving/assessing trustworthiness, and evaluation protocols for application-oriented workflows;
- Ethical considerations for decision-making based on computer vision-enabled analysis of geospatial imagery. What are bias sources, privacy/data protection, fairness, trust, accountability issues that are distinctive or particularly critical for computer vision models in the geospatial application domains? What constitutes suitable guardrails and what is still missing?
- Discussions on open challenges and opportunities from end-user, problem owners, and/or industry perspective. What should be the research community focusing on to augment or enable new CV4EO applications?
Format
The final program to include two keynote addresses, oral presentations, and a CV4EO community panel discussion. The planned format will consist of lightning talks + poster session, with the following manuscript submission types:
- Full research papers: presenting mature research on a specific problem or topic in the context of computer vision for Earth Observation data analysis. Submissions will be limited to 8 pages, including figures and tables, following the WACV style (same guidelines as the main conference). Papers will be peer-reviewed following the same WACV main conference policies (i.e., double blind) to ensure the quality and the clarity of the work. Accepted full research papers will be included in the Workshop Proceedings of WACV 2025
- Short/vision papers: short research articles or industry demonstrations of existing methods, toolkits, pipelines and best practices for CV4EO, as well as short papers discussing vision for future directions and/or statements on gaps and challenges for the development of AI technology and their applications in the geospatial domain. Accepted short papers will not be included in Workshop Proceedings of WACV 2024, but can be published in the CV4EO Workshop’s website upon author’s agreement. Moreover, a compilation with joint co-authorship of accepted short papers and main takeaways from the event is targeted for publication in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (JSTARS).
