36 datasets found

The dataset contains 2,000 frames, in which the first 800 frames for training and the remaining for testing.

The dataset contains 4,372 images, in which the first 800 frames for training and the remaining for testing.

The dataset contains 716 images, in which 400 images for training and the remaining for testing.

The dataset contains 482 images, in which 300 images for training and 182 images for testing.

Crowd counting has drawn increasing attention because of its fundamental role in social management. Six datasets are used in our experiments, i.e., GCC, ShanghaiTech PartA,...

Multi-modal crowd counting involves estimating crowd density from both visual and thermal/depth images.

Multi-modal crowd counting involves estimating crowd density from both visual and thermal/depth images.

Multi-modal crowd counting involves estimating crowd density from both visual and thermal/depth images.

A crowd counting dataset with over 1000 labeled videos captured by over 100 monitoring cameras.

Object detection on Unmanned Aerial Vehicles (UAVs) is still a challenging task. The recordings are mostly sparse and contain only small objects.

Crowd counting is a key aspect of crowd analysis and has been typically accomplished by estimating a crowd-density map and summing over the density values.

ShanghaiTechA and ShanghaiTechB are two crowd counting datasets containing images of crowds.

Crowd counting is a critical task in computer vision, with several important applications. However, existing counting methods rely on labor-intensive density map annotations,...

The CARPK dataset is a large-scale aerial dataset for counting cars in parking lots. It contains images with a top-down view covering 4 different parking lots.

The dataset contains 2,000 frames, in which the training set captures 601 to 1,400 and the testing set owns the remaining.

The dataset used in the paper for crowd counting, consisting of images with varying crowd densities.

The dataset used for crowd counting.

The dataset used in the paper for crowd counting, which includes five challenging benchmarks: ShanghaiTech Part A & Part B (SHHA & SHHB), UCF-QNRF (QNRF), JHU-Crowd++...

The Mall dataset is used for crowd counting, which contains 2000 annotated frames of over 60000 pedestrians with their head positions labeled.

The proposed E3D model is used for crowd counting, which captures the temporal dependencies in video sequences and combines global contextual information with local...