The V4AS workshop invites researchers and practitioners to submit original work addressing the challenges of computer vision under adverse weather and lighting conditions. Fog, rain, snow, ice, low light, nighttime, glare, and shadows all degrade the performance of sensors that power automated systems, from autonomous vehicles to surveillance networks. Submissions are encouraged across a broad range of topics, including image restoration and enhancement, robust feature extraction, domain adaptation for weather-affected data, sensor fusion strategies, and the development of benchmark datasets that reflect real-world conditions. Both theoretical contributions and applied systems are welcome, as the workshop aims to bridge the gap between laboratory performance and deployment in challenging outdoor environments.

A central theme of the call for papers is the gap between current algorithm performance on favorable, well-lit benchmark data and the severe degradation observed when the same methods encounter adverse conditions. Many state-of-the-art approaches assume clean inputs and fail gracefully when presented with rain streaks, dense fog, or nighttime illumination. The workshop seeks contributions that push beyond these limitations, whether through novel network architectures, physics-based modeling of weather effects, or data-driven approaches that leverage synthetic and real-world imagery. Authors are encouraged to discuss not only accuracy gains but also computational efficiency and practical robustness, ensuring that proposed solutions are viable for real-time deployment in autonomous systems and other safety-critical applications.

Beyond algorithmic contributions, the workshop welcomes position papers, survey work, and studies that critically examine existing benchmarks and evaluation protocols for weather-affected vision. Many current datasets are captured under controlled or simulated conditions, and their transferability to real-world scenarios remains an open question. Submissions may explore how to construct more representative test sets, how to measure robustness in a standardized way, or how to account for the wide variability of weather phenomena across seasons and geographic regions. The goal is to foster a community-wide conversation about evaluation practices, encouraging transparency and reproducibility so that progress in this field can be measured reliably and compared fairly across research groups.

The workshop also encourages interdisciplinary submissions that connect computer vision with related fields such as autonomous driving, robotics, remote sensing, and surveillance. Practical deployment scenarios often involve additional constraints, including limited computational resources, sensor calibration issues, and the need to operate continuously across changing conditions throughout the day and across seasons. Papers that address these systems-level concerns, such as online adaptation, uncertainty estimation, or fail-safe mechanisms, are particularly welcome. By bringing together researchers from academia and industry, the workshop aims to accelerate progress toward vision systems that maintain reliable performance regardless of the weather, ultimately supporting the safe and seamless operation of automated technologies in the real world.