12 datasets found

The proposed framework for analyzing medium dense crowd videos at various levels.

The UCSD dataset is a benchmark for the analysis of abnormal detection and localization in crowded scenes.

ShanghaiTech PartA and PartB datasets for crowd counting and localization.

The UMN dataset consists of three different crowd scenes, and the dataset has 11 videos from these scenes, with a resolution of 240 × 320. Each video sequence represents a...

The dataset contains images of crowds with varying densities and sizes.

The proposed FIDT map is a non-overlap map, which utilizes local maxima to represent the head's center.

The UCF CC 50 dataset contains 50 images collected from publicly available web images.

ShanghaiTech Part B is a crowd counting dataset that contains 400 training images and 316 test images.

ShanghaiTech Part A is a crowd counting dataset that contains 300 training images and 182 test images.

A video crowd counting dataset, which consists of 13 different scenes, 100 image sequences, 150,000 frames.

Crowd localization is a new computer vision task, evolved from crowd counting. Different from the latter, it provides more precise location information for each instance, not...

The ShanghaiTech dataset includes 330 training and 107 test videos recorded at 13 different background locations.