Multi-source Multi-scale Counting in Extremely Dense Crowd Images

By Haroon Idrees, Imran Saleemi, Cody Seibert, Mubarak Shah. 2013

1,019 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

DeepSetNet: Predicting Sets with Deep Neural Networks (2016, background): “It is important to bear in mind that unlike many existing approaches that learn to count [1, 5, 12, 20, 24, 35, 50, 51], our main goal is not object counting.”
CrowdNet: A Deep Convolutional Network for Dense Crowd Counting (2016, methodology): “Mean Absolute Error Learning to Count [12] 493.4 Density-aware Detection [16] 655.7 FHSc [8] 468.0 Cross-Scene Counting [19] 467.0 Proposed 452.5 Table 1: Quantitative results of our approach along with other state-of-the-art methods on UCF CC 50 Dataset.”
Crowd Counting and Density Estimation by Trellis Encoder-Decoder Networks (2019, methodology): “Most crowd counting methods are based on detection [18, 9, 45], regression [28, 4, 10], and density estimation [17, 53, 2, 11].”
CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes (2018, methodology): “5-fold cross-validation is performed following the standard setting in [22].”
CNN-Based cascaded multi-task learning of high-level prior and density estimation for crowd counting (2017, methodology): “resentations and catered only to low density crowded scenes [10].”
Iterative Crowd Counting (2018, methodology): “[6] 419.5 487.1 Crowd CNN [19] 467.0 498.5 Crowdnet [1] 452.5 - MCNN [20] 377.6 509.1 Hydra2s [13] 333.7 425.6 Switch CNN [15] 318.1 439.2 CP-CNN [16] 295.8 320.9 ic-CNN (proposed) 260.9 365.5”
ADCrowdNet: An Attention-Injective Deformable Convolutional Network for Crowd Understanding (2018, methodology): “We used 5-fold-crossvalidation setting described in [12].”
Interactive Object Counting (2014, background): “Towards this goal, several computer vision approaches that use machine learning in order to automate such tasks have been suggested [5, 8, 10, 12].”
Switching Convolutional Neural Network for Crowd Counting (2017, methodology): “• We evidence state-of-the-art performance on all major crowd counting datasets including ShanghaiTech dataset [19], UCF CC 50 dataset [6] and WorldExpo’10 dataset [18].”
Learning From Synthetic Data for Crowd Counting in the Wild (2019, methodology): “Dataset Number Average Count Statistics of Images Resolution Total Min Ave Max UCSD [8] 2,000 158× 238 49,885 11 25 46 Mall [9] 2,000 480× 640 62,325 13 31 53 UCF CC 50 [14] 50 2101× 2888 63,974 94 1,279 4,543 WorldExpo’10[41] 3,980 576× 720 199,923 1 50 253 SHT A [43] 482 589× 868 241,677 33 501 3,”
Deep People Counting in Extremely Dense Crowds (2015, methodology): “We display ground truth, estimation made by [6] and estimation made by our method in black, blue and red box, separately.”
Scale Aggregation Network for Accurate and Efficient Crowd Counting (2018, methodology): “The UCF CC 50 dataset [21] includes 50 annotated crowd images.”
A Thermal Infrared Video Benchmark for Visual Analysis (2014, methodology): “Fortunately, techniques to count the objects without using an object detector exist [17], [18], [19].”
Crowd Counting via Adversarial Cross-Scale Consistency Pursuit (2018, methodology): “[8] which is a very challenging dataset composed of 50 annotated crowd images with a large variance in crowd counts and scenes.”
Towards Perspective-Free Object Counting with Deep Learning (2016, methodology): “Two have been proposed for the crowd counting problem: the UCSD pedestrian [4] and the UCF CC 50 [9] datasets.”
Context-Aware Crowd Counting (2018, methodology): “[14] 315 508 MCNN [42] 277 426 Encoder-Decoder [3] 270 478 CMTL [35] 252 514 Switch-CNN [31] 228 445 Resnet101 [12] 190 277 Densenet201 [13] 163 226 Idrees et al.”
Distribution Matching for Crowd Counting (2020, methodology): “In all existing crowd counting datasets [15, 60, 14, 51], the annotated density map for each training image is a sparse binary mask, where each individual person is marked with a single dot on their head or forehead.”
Rethinking Counting and Localization in Crowds: A Purely Point-Based Framework (2021, methodology): “Specifically, extensive experiments are conducted on four challenging datasets, including ShanghaiTech PartA and PartB [41], UCF CC 50 [9], UCF-QNRF [10] and NWPU-Crowd [35].”