Congested Crowd Counting via Adaptive Multi-Scale Context Learning

Zhang, Yani; Zhao, Huijuan; Duan, Zhenhai; Huang, Liangjun; Deng, Jiahao; Zhang, Qing

doi:10.3390/s21113777

Cited by 13 publications

(4 citation statements)

References 93 publications

(124 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This research served as a source of inspiration for our work. In addition, some recent works [26,31,32] have shown promising results and contributed to the advancement of crowd counting techniques.…”

Section: Density Map-based Methodsmentioning

confidence: 99%

Foreground Segmentation-Based Density Grading Networks for Crowd Counting

Liu,

Zhou,

Zhou

et al. 2023

Sensors

View full text Add to dashboard Cite

Estimating object counts within a single image or video frame represents a challenging yet pivotal task in the field of computer vision. Its increasing significance arises from its versatile applications across various domains, including public safety and urban planning. Among the various object counting tasks, crowd counting is particularly notable for its critical role in social security and urban planning. However, intricate backgrounds in images often lead to misidentifications, wherein the complex background is mistaken as the foreground, thereby inflating forecasting errors. Additionally, the uneven distribution of crowd density within the foreground further exacerbates predictive errors of the network. This paper introduces a novel architecture with a three-branch structure aimed at synergistically incorporating hierarchical foreground information and global scale information into density map estimation, thereby achieving more precise counting results. Hierarchical foreground information guides the network to perform distinct operations on regions with varying densities, while global scale information evaluates the overall density level of the image and adjusts the model’s global predictions accordingly. We also systematically investigate and compare three potential locations for integrating hierarchical foreground information into the density estimation network, ultimately determining the most effective placement.Through extensive comparative experiments across three datasets, we demonstrate the superior performance of our proposed method.

show abstract

Section: Density Map-based Methodsmentioning

confidence: 99%

Foreground Segmentation-Based Density Grading Networks for Crowd Counting

Liu,

Zhou,

Zhou

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Compared to the PPAM block in our model, our method takes fewer computations to process scaleaware information and the Swin-Transformer block is more effective in extracting global context information. Tian et al [41] and Zhang et al [42] put forward guidance branch to their lightweight model to learn localization information in their work; such a technique needs precise head location coordinates to guide location task and that is not mandatory in lightweight crowd-counting tasks. When it comes to hybrid network architecture, Sun et al [43] introduce Transformer blocks after each downscale convolution block to separately model scale-varied information stage by stage; it is not computationally efficient to introduce multiple Transformer blocks for the same thing.…”

Section: Lightweight Crowd-counting Modelsmentioning

confidence: 99%

A Weakly Supervised Hybrid Lightweight Network for Efficient Crowd Counting

Chen,

Zhao,

Gao

et al. 2024

Electronics

Self Cite

View full text Add to dashboard Cite

Crowd-counting networks have become the mainstream method to deploy crowd-counting techniques on resource-constrained devices. Significant progress has been made in this field, with many outstanding lightweight models being proposed successively. However, challenges like scare-variation, global feature extraction, and fine-grained head annotation requirements still exist in relevant tasks, necessitating further improvement. In this article, we propose a weakly supervised hybrid lightweight crowd-counting network that integrates the initial layers of GhostNet as the backbone to efficiently extract local features and enrich intermediate features. The incorporation of a modified Swin-Transformer block addresses the need for effective global context information. A Pyramid Pooling Aggregation Module handles the inherent scale variation problem in crowd-counting tasks in a more computation-efficient way. This module, along with the cross-attention module, serves as bridges to promote the feature information flow between local features and global context information. Finally, a simplified regressor module is designed to allow the proposed model with weakly supervised guidance for training to avoid precise location-level annotations, noting that the omission of density map generation makes the proposed network more lightweight. Our results on the UCF-QNRF dataset indicate our model is 8.73% and 12.17% more accurate on MAE and MSE metrics, respectively, than the second-best ARNet, in which the parameters decrease by 4.52%. On the ShanghaiTech A dataset, MAE and MSE drop 1.5% and 3.2%, respectively, compared to the second-best PDDNet. The experimental results for accuracy and inference speed evaluation on some mainstream datasets validate the effective design principle of our model.

show abstract

“…Through a multi-task learning method, the scale variation problem was solved and crowd counting was achieved accurately. Through applying an edge computing method, Wang et al [22] achieved crowd density estimation efficiently to solve the problem of high network latency caused by the deployment of the density estimation platform in the server; Zhang et al [23] proposed the adaptive multi-scale context aggregation network (MSCANet) to obtain the full-scale information of the crowd. After the information of different scales was extracted by the network, this information was fused by the network adaptively, which was suitable for crowd counting.…”

Section: Related Workmentioning

confidence: 99%

Advanced Pedestrian State Sensing Method for Automated Patrol Vehicle Based on Multi-Sensor Fusion

Wang

Cheng

Wang

et al. 2022

Sensors

View full text Add to dashboard Cite

At present, the COVID-19 pandemic still presents with outbreaks occasionally, and pedestrians in public areas are at risk of being infected by the viruses. In order to reduce the risk of cross-infection, an advanced pedestrian state sensing method for automated patrol vehicles based on multi-sensor fusion is proposed to sense pedestrian state. Firstly, the pedestrian data output by the Euclidean clustering algorithm and the YOLO V4 network are obtained, and a decision-level fusion method is adopted to improve the accuracy of pedestrian detection. Then, combined with the pedestrian detection results, we calculate the crowd density distribution based on multi-layer fusion and estimate the crowd density in the scenario according to the density distribution. In addition, once the crowd aggregates, the body temperature of the aggregated crowd is detected by a thermal infrared camera. Finally, based on the proposed method, an experiment with an automated patrol vehicle is designed to verify the accuracy and feasibility. The experimental results have shown that the mean accuracy of pedestrian detection is increased by 17.1% compared with using a single sensor. The area of crowd aggregation is divided, and the mean error of the crowd density estimation is 3.74%. The maximum error between the body temperature detection results and thermometer measurement results is less than 0.8°, and the abnormal temperature targets can be determined in the scenario, which can provide an efficient advanced pedestrian state sensing technique for the prevention and control area of an epidemic.

show abstract

Congested Crowd Counting via Adaptive Multi-Scale Context Learning

Cited by 13 publications

References 93 publications

Foreground Segmentation-Based Density Grading Networks for Crowd Counting

Foreground Segmentation-Based Density Grading Networks for Crowd Counting

A Weakly Supervised Hybrid Lightweight Network for Efficient Crowd Counting

Advanced Pedestrian State Sensing Method for Automated Patrol Vehicle Based on Multi-Sensor Fusion

Contact Info

Product

Resources

About