In many animal species, courtship-related social behavior is essential for mating success. Given the diversity and complexity of animal courtship-related social behavior, it is challenging to automatically recognize courtship-related social behavior. Zebrafish has emerged as an important animal model for studying courtship behavior. However, to the best of our knowledge, no research has been conducted on the automated recognition of zebrafish courtship-related social behavior. Therefore, this work builds a benchmark dataset of zebrafish courtship-related social behavior and proposes a zebrafish courtship-related social behavior automated recognition (ZCSBAR) framework. The ZCSBAR framework consists of three functional modules, including pose estimation, long-time identity tracking and social behavior recognition. The experimental results show that the ZCSBAR achieves high performance in zebrafish social behavior recognition, specifically the identity recognition index IDF1 reaches 0.902 and the average behavior recognition accuracy reaches 0.859. These findings highlight the potential of ZCSBAR for quantifying the courtship-related social behavior in the fields of animal behavior and reproduction.

This comprehensive study investigates the application of Large Language Models (LLMs) for detecting and protecting Personally Identifiable Information (PII) in archival data, a pressing concern for archives under the mandate to increase public access while safeguarding personal privacy. The paper juxtaposes traditional supervised learning methods against LLMs’ unsupervised capabilities in PII detection, unveiling LLMs as viable alternatives capable of achieving satisfactory performance levels without the need for extensive training datasets. Through empirical analysis, the study validates the feasibility of LLMs in identifying sensitive information within large volumes of archival material. The findings highlight LLMs’ significant interpretability, providing understandable rationale behind PII identification—a feature that not only enhances trust in AI applications but also aids archival staff in the review process. This research contributes novel insights into the intersection of AI and archival science, presenting LLMs as powerful tools for addressing the twin challenges of data accessibility and privacy.

We present PANDA, the first gigaPixel-level humAN-centric viDeo dAtaset, for large-scale, long-term, and multi-object visual analysis. The videos in PANDA were captured by a gigapixel camera and cover real-world scenes with both wide field-of-view (~1 square kilometer area) and high-resolution details (~gigapixel-level/frame). The scenes may contain 4k head counts with over 100× scale variation. PANDA provides enriched and hierarchical ground-truth annotations, including 15,974.6k bounding boxes, 111.8k fine-grained attribute labels, 12.7k trajectories, 2.2k groups and 2.9k interactions. We benchmark the human detection and tracking tasks. Due to the vast variance of pedestrian pose, scale, occlusion and trajectory, existing approaches are challenged by both accuracy and efficiency. Given the uniqueness of PANDA with both wide FoV and high resolution, a new task of interaction-aware group detection is introduced. We design a `global-to-local zoom-in' framework, where global trajectories and local interactions are simultaneously encoded, yielding promising results. We believe PANDA will contribute to the community of artificial intelligence and praxeology by understanding human behaviors and interactions in large-scale real-world scenes. PANDA Website: http://www.panda-dataset.com.

In this study, a new method of extracting urban extents using relation among statistical features and decision rules was developed. Results indicate that the data from different years produced the same decision rules. The experimental results also demonstrate that the method can achieve better results than the global threshold method. A main advantage of the proposed method is that the rules derived from data of a specific year can be transferred to the data of other years, which makes it promising for rapidly monitoring urban areas regionally using DMSP/OLS NTL time series data. However, the proposed method needs further validation using more data in different geographical regions.