AVSS: a new benchmark for airport video semantic segmentation
Keywords: Airport Ground, Semantic Segmentation, Video Surveillance
Abstract: Airport video semantic segmentation is fundamental to airport surveillance applications, yet there currently lacks a specialized benchmark and algorithms for this task. In this paper, we introduce the first large-scale Airport Video Semantic Segmentation dataset (AVSS) for airport surveillance. AVSS comprises 18 common semantic categories at airports, and 250 videos, totaling over 140,000 frames with accurate manual annotations. AVSS covers a wide range of challenges for airport video surveillance, such as extreme multi-scale, intra-class diversity, inter-class similarity, etc. We analyze statistical information and evaluate 17 state-of-the-art (SOTA) semantic segmentation algorithms on AVSS. The significant performance degradation indicates that current models are far from practical application. Furthermore, we discuss how to develop video semantic segmentation algorithms for airport surveillance and the generalizability of AVSS to other tasks and datasets. AVSS serves as a research resource for airport semantic segmentation and a robustness evaluation tool for segmentation algorithms in practical applications. AVSS is available at www.agvs-caac.com/avss/avss.html.
Primary Area: datasets and benchmarks
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 8827
Loading