Going Beyond the Regression Paradigm with Accurate Dot Prediction for Dense Crowds

Sam, Deepak Babu; Peri, Skand Vishwanath; Mukuntha, N S; Babu, R. Venkatesh

doi:10.1109/wacv45572.2020.9093386

Cited by 11 publications

(9 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [1], bounding boxes around point annotations were initialized using their distance to the nearest neighbour head and while training a CNN to output box annotations, updated bounding boxes around the heads to obtain the most suitable size in the anchor box set for the corresponding head. In [20], two scales of the input image feature maps (one-fourth and one-eighth) were used depending on the sparsity of the crowd to output a point map. Crowd counting was also cast as a classification task of dot prediction with point supervision, dropping the prevalent density regression.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Multiscale Crowd Counting and Localization By Multitask Point Supervision

Zand¹,

Haleh²,

Farley³

et al. 2022

Preprint

View full text Add to dashboard Cite

We propose a multitask approach for crowd counting and person localization in a unified framework. As the detection and localization tasks are well-correlated and can be jointly tackled, our model benefits from a multitask solution by learning multiscale representations of encoded crowd images, and subsequently fusing them. In contrast to the relatively more popular density-based methods, our model uses point supervision to allow for crowd locations to be accurately identified. We test our model on two popular crowd counting datasets, ShanghaiTech A and B, and demonstrate that our method achieves strong results on both counting and localization tasks, with MSE measures of 110.7 and 15.0 for crowd counting and AP measures of 0.71 and 0.75 for localization, on ShanghaiTech A and B respectively. Our detailed ablation experiments show the impact of our multiscale approach as well as the effectiveness of the fusion

show abstract

Section: Related Workmentioning

confidence: 99%

“…A threshold was then applied on this map to generate the final accurate dot predictions. Therefore in [20], the fusion module was not learnable and the count highly depended on the accuracy of the threshold value which might be different across different images.…”

Section: Related Workmentioning

confidence: 99%

Multiscale Crowd Counting and Localization By Multitask Point Supervision

Zand¹,

Haleh²,

Farley³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Liu et al [12] propose a curriculum learning strategy to generate pseudo box-level labels from point annotations. Sam et al [35] train a pixel-wise binary classifier to detect people instead of regressing local crowd density, which proposes a novel multi-scale architecture incorporating top-down feedback to address the scale variation. Furthermore, they propose the upgraded version, LSC-CNN [13], which is a tailor-made dense object detection method that predicts position, size for each head simultaneously only with point-level annotations.…”

Section: Crowd Localizationmentioning

confidence: 99%

LDC-Net: A Unified Framework for Localization, Detection and Counting in Dense Crowds

wang,

Han,

Gao

et al. 2021

Preprint

View full text Add to dashboard Cite

The rapid development in visual crowd analysis shows a trend to count people by positioning or even detecting, rather than simply summing a density map. It also enlightens us back to the essence of the field, detection to count, which can give more abundant crowd information and has more practical applications. However, some recent work on crowd localization and detection has two limitations: 1) The typical detection methods can not handle the dense crowds and a large variation in scale; 2) The density map heuristic methods suffer from performance deficiency in position and box prediction, especially in high density or large-size crowds. In this paper, we devise a tailored baseline for dense crowds location, detection, and counting from a new perspective, named as LDC-Net for convenience, which has the following features: 1) A strong but minimalist paradigm to detect objects by only predicting a location map and a size map, which endows an ability to detect in a scene with any capacity (0 ∼ 10, 000+ persons); 2) Excellent cross-scale ability in facing a large variation, such as the head ranging in 0 ∼ 100, 000+ pixels; 3) Achieve superior performance in location and box prediction tasks, as well as a competitive counting performance compared with the density-based methods. Finally, the source code and pre-trained models will be released.

show abstract

“…Recently, crowd localization methods have been developed for dense crowds, like [2,17,18], showing that those methods can localize crowds as well as obtain a comparable counting result. The localization process is to produce the keypoints [2,19], blobs [17], or bounding boxes [18,20,21] on each head (see Fig. 1 (a)), and 978-1-6654-4989-2/21/$31.00 ©2021 IEEE Fig.…”

Section: (B) Congested Scene Recognition Networkmentioning

confidence: 99%

Dense Point Prediction: A Simple Baseline for Crowd Counting and Localization

Wang

Hou

Chau

2021

2021 IEEE International Conference on Multimedia &Amp; Expo Workshops (ICMEW)

View full text Add to dashboard Cite

In this paper, we propose a simple yet effective crowd counting and localization network named SCALNet. Unlike most existing works that separate the counting and localization tasks, we consider those tasks as a pixel-wise dense prediction problem and integrate them into an end-to-end framework. Specifically, for crowd counting, we adopt a counting head supervised by the Mean Square Error (MSE) loss. For crowd localization, the key insight is to recognize the keypoint of people, i.e., the center point of heads. We propose a localization head to distinguish dense crowds trained by two loss functions, i.e., Negative-Suppressed Focal (NSF) loss and False-Positive (FP) loss, which balances the positive/negative examples and handles the false-positive predictions. Experiments on the recent and large-scale benchmark, NWPU-Crowd, show that our approach outperforms the state-of-the-art methods by more than 5% and 10% improvement in crowd localization and counting tasks, respectively.

show abstract

Going Beyond the Regression Paradigm with Accurate Dot Prediction for Dense Crowds

Cited by 11 publications

References 22 publications

Multiscale Crowd Counting and Localization By Multitask Point Supervision

Multiscale Crowd Counting and Localization By Multitask Point Supervision

LDC-Net: A Unified Framework for Localization, Detection and Counting in Dense Crowds

Dense Point Prediction: A Simple Baseline for Crowd Counting and Localization

Contact Info

Product

Resources

About