CVC-14

数据集由两组序列组成。这些序列被命名为白天和夜晚集，指的是它们被获取的一天中的时刻，并且可视和FIR取决于用户记录序列的相机。用于训练白天的3695 imegas，晚上的3390图像，为每个序列注释大约1500个强制性行人。用于测试两个序列的700图像，其中白天约2000行人，晚上约1500行人。

The dataset comprises two sequence groups, namely the Daytime set and the Nighttime set, which are named based on the time of day when the data was acquired. The visibility and FIR characteristics are dependent on the camera used by the user to record the sequences. For the training subset, the daytime set contains 3695 images and the nighttime set includes 3390 images, with approximately 1500 mandatory pedestrian annotations per sequence. For the test subset, there are 700 images across both sets, with roughly 2000 pedestrians in the daytime portion and 1500 pedestrians in the nighttime portion.