Coarse to Fine: Domain Adaptive Crowd Counting via Adversarial Scoring Network

Zou, Zhikang; Qu, Xiaoye; Zhou, Pan; Xu, Shuangjie; Ye, Xiaoqing; Wu, Wenhao; Ye, Jin

doi:10.1145/3474085.3475377

Cited by 26 publications

(13 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 3 (* means that only the source domain image passes the style transfer layer when training the model) shows that our method gets the best MAE metric. Particularly, the RMSE of our method improves by 7.6% compared with the recent domain adaptation method ASNet [10]. And from the results in the table, we can conclude that our method gets a better result than the results that only let the source domain data pass through the style transfer layer.…”

Section: Comparison With Other State-of-the-art Algorithmsmentioning

confidence: 72%

“…Recently, to address different crowd scales and density distributions between different domains, CODA [8] performed adversarial training with multi‐scale image pyramids from two domains and achieved results close to the state‐of‐the‐art fully supervised model trained on the target domain. ASNet [10] adopted a dual discriminators strategy to close the source domain to the target domain from the perspective of global and local feature spaces through adversarial learning. However, these methods over‐optimise samples close to the target domain on the source domain, so the model's performance is improved on the target domains but significantly degraded on the source domains.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Domain adaptive crowd counting via dynamic scale aggregation network

Huo

Wang

Qiao

et al. 2023

IET Computer Vision

View full text Add to dashboard Cite

Crowd counting is an important research topic in computer vision. Its goal is to estimate the people's number in an image. Researchers have dramatically improved counting accuracy in recent years by regressing density maps. However, because of the inherent domain shift, the model trained on an expensive manually labelled dataset (source domain) does not perform well on a dataset with scarce labels (target domain). For this issue, a novel dynamic scale aggregation network (DSANet) is proposed to reduce the gaps in style and cross‐domain head scale variations. Specifically, a practical style transfer layer is introduced to reduce the appearance discrepancy between the source and target domains. Then, the translated source and target domain samples are encoded by a generator consisting of the VGG16 network and the dynamic scale aggregation modules (DSA Modules) and produce corresponding density maps. The DSA module can adaptively adjust parameters according to the input features and effectively fuse multi‐scale information to overcome the cross‐domain head scale variations. Next, a discriminator judges the input density map from the source or target domain. Last, domain distributions are aligned through adversarial between the generator and the discriminator. The experiments show that our network outperforms the current state‐of‐the‐art methods and can improve the target domain's performance while maintaining the source domain's performance without significant degradation.

show abstract

Section: Comparison With Other State-of-the-art Algorithmsmentioning

confidence: 72%

Section: Introductionmentioning

confidence: 99%

Domain adaptive crowd counting via dynamic scale aggregation network

Huo

Wang

Qiao

et al. 2023

IET Computer Vision

View full text Add to dashboard Cite

show abstract

“…Zou et al [3] proposed an adversarial scoring network (ASNet) that gradually closes the cross-domain gap from coarse-to fine-grained. The coarse-grained phase designs a dual discriminator strategy to adapt the source domain from the global and local feature space perspectives close to the target domain.…”

Section: Methods Based On the Distribution Strategymentioning

confidence: 99%

A Survey of Crowd Counting Algorithm Based on Domain Adaptation

Wang¹,

Luo²

2023

AJST

View full text Add to dashboard Cite

Crowd counting, the task of estimating the number of individuals in a crowded scene, has gained increasing attention in computer vision research. However, crowd counting remains a challenging problem due to the complex and diverse nature of crowd scenes. In recent years, domain adaptation has emerged as a promising approach to improve crowd counting performance by adapting a pre-trained model to a target domain with different characteristics. This paper provides a survey of domain adaptation-based crowd counting algorithms, including their methods, datasets, and evaluation metrics. Overall, domain adaptation shows great potential in improving the accuracy and robustness of crowd counting algorithms, and further research in this direction is expected to lead to more effective and practical crowd counting solutions.

show abstract

“…Cross-domain / Multi-domain Learning. Many researchers exploit the cross-domain problems [40,41,42,43,25] in crowd counting, including cross-scene [32], cross-view [44], cross-modal [45], etc. Adversarial Scoring Network [41] is applied to adapt to the target domain from coarse to fine granularity.…”

Section: Related Workmentioning

confidence: 99%

“…Many researchers exploit the cross-domain problems [40,41,42,43,25] in crowd counting, including cross-scene [32], cross-view [44], cross-modal [45], etc. Adversarial Scoring Network [41] is applied to adapt to the target domain from coarse to fine granularity. Besides, cross-domain features can be extracted by the message-passing mechanisms based on a graph neural network [46].…”

Section: Related Workmentioning

confidence: 99%

Forget Less, Count Better: A Domain-Incremental Self-Distillation Learning Benchmark for Lifelong Crowd Counting

Gao¹,

Li²,

Shan³

et al. 2022

Preprint

View full text Add to dashboard Cite

Crowd Counting has important applications in public safety and pandemic control. A robust and practical crowd counting system has to be capable of continuously learning with the new-coming domain data in real-world scenarios instead of fitting one domain only. Off-the-shelf methods have some drawbacks to handle multiple domains. 1) The models will achieve limited performance (even drop dramatically) among old domains after training images from new domains due to the discrepancies of intrinsic data distributions from various domains, which is called catastrophic forgetting. 2) The well-trained model in a specific domain achieves imperfect performance among other unseen domains because of the domain shift. 3) It leads to linearly-increased storage overhead either mixing all the data for training or simply training dozens of separate models for different domains when new ones are available. To overcome these issues, we investigate a new task of crowd counting under the incremental domains training setting, namely, Lifelong Crowd Counting. It aims at alleviating the catastrophic forgetting and improving the generalization ability using a single model updated by the incremental domains. To be more specific, we propose a selfdistillation learning framework as a benchmark (Forget Less, Count Better, FLCB) for lifelong crowd counting, which helps the model sustainably leverage previous meaningful knowledge for better crowd counting to mitigate the forgetting when the new data arrive. Meanwhile, a new quantitative metric, normalized backward transfer (nBwT), is developed to evaluate the forgetting degree of the model in the lifelong learning process. Extensive experimental results demonstrate the superiority of our proposed benchmark in achieving a low catastrophic forgetting degree and strong generalization ability.

show abstract

Coarse to Fine: Domain Adaptive Crowd Counting via Adversarial Scoring Network

Cited by 26 publications

References 49 publications

Domain adaptive crowd counting via dynamic scale aggregation network

Domain adaptive crowd counting via dynamic scale aggregation network

A Survey of Crowd Counting Algorithm Based on Domain Adaptation

Forget Less, Count Better: A Domain-Incremental Self-Distillation Learning Benchmark for Lifelong Crowd Counting

Contact Info

Product

Resources

About