Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-Learning
Counting people in crowd scenarios is extensively conducted in drone inspections, video surveillance, and public safety applications. Today, crowd count algorithms with supervised learning have improved significantly, but with a reliance on a large amount of manual annotation. However, in real world...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-12-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/11/24/12037 |
| _version_ | 1797506744269668352 |
|---|---|
| author | Xiaoyu Hou Jihui Xu Jinming Wu Huaiyu Xu |
| author_facet | Xiaoyu Hou Jihui Xu Jinming Wu Huaiyu Xu |
| author_sort | Xiaoyu Hou |
| collection | DOAJ |
| description | Counting people in crowd scenarios is extensively conducted in drone inspections, video surveillance, and public safety applications. Today, crowd count algorithms with supervised learning have improved significantly, but with a reliance on a large amount of manual annotation. However, in real world scenarios, different photo angles, exposures, location heights, complex backgrounds, and limited annotation data lead to supervised learning methods not working satisfactorily, plus many of them suffer from overfitting problems. To address the above issues, we focus on training synthetic crowd data and investigate how to transfer information to real-world datasets while reducing the need for manual annotation. CNN-based crowd-counting algorithms usually consist of feature extraction, density estimation, and count regression. To improve the domain adaptation in feature extraction, we propose an adaptive domain-invariant feature extracting module. Meanwhile, after taking inspiration from recent innovative meta-learning, we present a dynamic-<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>β</mi></semantics></math></inline-formula> MAML algorithm to generate a density map in unseen novel scenes and render the density estimation model more universal. Finally, we use a counting map refiner to optimize the coarse density map transformation into a fine density map and then regress the crowd number. Extensive experiments show that our proposed domain adaptation- and model-generalization methods can effectively suppress domain gaps and produce elaborate density maps in cross-domain crowd-counting scenarios. We demonstrate that the proposals in our paper outperform current state-of-the-art techniques. |
| first_indexed | 2024-03-10T04:37:11Z |
| format | Article |
| id | doaj.art-f633231dcbb14497b531c0fca822d898 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T04:37:11Z |
| publishDate | 2021-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-f633231dcbb14497b531c0fca822d8982023-11-23T03:42:14ZengMDPI AGApplied Sciences2076-34172021-12-0111241203710.3390/app112412037Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-LearningXiaoyu Hou0Jihui Xu1Jinming Wu2Huaiyu Xu3Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, ChinaShanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, ChinaShanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, ChinaShanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, ChinaCounting people in crowd scenarios is extensively conducted in drone inspections, video surveillance, and public safety applications. Today, crowd count algorithms with supervised learning have improved significantly, but with a reliance on a large amount of manual annotation. However, in real world scenarios, different photo angles, exposures, location heights, complex backgrounds, and limited annotation data lead to supervised learning methods not working satisfactorily, plus many of them suffer from overfitting problems. To address the above issues, we focus on training synthetic crowd data and investigate how to transfer information to real-world datasets while reducing the need for manual annotation. CNN-based crowd-counting algorithms usually consist of feature extraction, density estimation, and count regression. To improve the domain adaptation in feature extraction, we propose an adaptive domain-invariant feature extracting module. Meanwhile, after taking inspiration from recent innovative meta-learning, we present a dynamic-<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>β</mi></semantics></math></inline-formula> MAML algorithm to generate a density map in unseen novel scenes and render the density estimation model more universal. Finally, we use a counting map refiner to optimize the coarse density map transformation into a fine density map and then regress the crowd number. Extensive experiments show that our proposed domain adaptation- and model-generalization methods can effectively suppress domain gaps and produce elaborate density maps in cross-domain crowd-counting scenarios. We demonstrate that the proposals in our paper outperform current state-of-the-art techniques.https://www.mdpi.com/2076-3417/11/24/12037crowd countingdomain adaptationcross-domainmeta-learningsynthetic dataset |
| spellingShingle | Xiaoyu Hou Jihui Xu Jinming Wu Huaiyu Xu Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-Learning Applied Sciences crowd counting domain adaptation cross-domain meta-learning synthetic dataset |
| title | Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-Learning |
| title_full | Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-Learning |
| title_fullStr | Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-Learning |
| title_full_unstemmed | Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-Learning |
| title_short | Cross Domain Adaptation of Crowd Counting with Model-Agnostic Meta-Learning |
| title_sort | cross domain adaptation of crowd counting with model agnostic meta learning |
| topic | crowd counting domain adaptation cross-domain meta-learning synthetic dataset |
| url | https://www.mdpi.com/2076-3417/11/24/12037 |
| work_keys_str_mv | AT xiaoyuhou crossdomainadaptationofcrowdcountingwithmodelagnosticmetalearning AT jihuixu crossdomainadaptationofcrowdcountingwithmodelagnosticmetalearning AT jinmingwu crossdomainadaptationofcrowdcountingwithmodelagnosticmetalearning AT huaiyuxu crossdomainadaptationofcrowdcountingwithmodelagnosticmetalearning |