A subset of 5000 frames of fish schooling data from "Inferring the structure and dynamics of interactions in schooling fish" by Katz et al.

The data is a JSON format file containing the position, velocity, and fish identifier data for 300 golden shiners in a shallow water tank. There are 5000 individual frames (samples of position and velocity) corresponding to video taken at a rate of 30 frames/s and analysed to extract individual fish’s trajectories. The fields px, py, vx, vy correspond to the x- and y-components of each detected fish’s position and velocity. The onfish field gives each fish in the frame a unique identifier. When an individual can no longer be distinguished, its identifier is retired; once the fish is again being tracked it is assigned a new ID. This data is a subsample of the frames acquired by the creators in the study https://doi.org/10.1073/pnas.1107583108, refer to it and its supplementary information for more details.

本数据集为JSON格式文件，收录了浅水槽中300条金色鲦鱼的位置、速度及个体标识数据。数据集包含5000帧独立采样帧（包含位置与速度采样数据），对应拍摄帧率为30帧/秒的视频，经分析处理提取得到各受试个体的运动轨迹。字段px、py、vx、vy分别对应每条被检测鱼类的位置与速度的x、y分量。onfish字段为每一帧中的每条鱼分配唯一识别标识。当某条个体无法被识别追踪时，其原有标识将被注销；当该个体重新被追踪到时，将被分配新的标识ID。本数据集为该研究创作者所采集帧的子样本，相关研究的DOI为https://doi.org/10.1073/pnas.1107583108，如需获取更多细节，请参阅该文献及其补充材料。