相关参考文献
注意:仅列出部分参考文献,下载原文获取全部文献信息。
Article
Computer Science, Information Systems
Siddharth Singh Savner et al.
Summary: Convolutional neural networks (CNNs) have been dominant in the field of computer vision for a long time, but they lack the ability to model the global context due to their limited receptive field. Transformers, as attention-based architectures, excel at capturing global context. However, there is limited research on the effectiveness of transformers in crowd counting, and the existing crowd-counting methods relying on density map regression are laborious and error-prone. To address these issues, this paper proposes a weakly-supervised crowd counting method using a pyramid vision transformer, which achieves state-of-the-art performance and remarkable generalizability.
JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION
(2023)
Article
Computer Science, Artificial Intelligence
Mingjie Wang et al.
Summary: This paper proposes a novel and efficient counting model, CrowdMLP, which regresses total counts by designing a multi-granularity MLP regressor that models global dependencies of embeddings. The model uses a locally-focused pre-trained frontend to extract crude feature maps with spatial cues and tokenizes the crude embeddings and raw crowd scenes at different granularities. The study also introduces a self-supervised proxy task, Split-Counting, to overcome limited samples and the lack of spatial hints.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Information Systems
Dingkang Liang et al.
Summary: This paper proposes TransCrowd, a weakly-supervised crowd counting method based on transformers. By utilizing the self-attention mechanism of transformers, TransCrowd effectively extracts semantic crowd information, addressing the limited receptive fields for context modeling in traditional CNN methods. Experiments show that TransCrowd outperforms other weakly-supervised CNN methods and achieves competitive performance compared to some fully-supervised counting methods.
SCIENCE CHINA-INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Junyu Gao et al.
Summary: This paper focuses on achieving precise instance localization in high-density crowd scenes by proposing a Dilated Convolutional Swin Transformer (DCST) and introducing a window-based vision transformer. Experimental results demonstrate the effectiveness of the proposed methods.
Proceedings Paper
Computer Science, Artificial Intelligence
Dingkang Liang et al.
Summary: In this paper, an elegant and end-to-end Crowd Localization Transformer (CLTR) is proposed to solve the task of crowd localization. The proposed method treats crowd localization as a direct set prediction problem and introduces a KMO-based Hungarian matcher to reduce ambiguous points and generate more reasonable matching results. Extensive experiments demonstrate the effectiveness of the proposed method.
COMPUTER VISION - ECCV 2022, PT I
(2022)
Article
Computer Science, Artificial Intelligence
Yinjie Lei et al.
Summary: This paper focuses on weakly-supervised crowd counting, where a model is learned from a small amount of location-level annotations and a large amount of count-level annotations. The study reveals that directly regressing the integral of density maps to object count is not satisfactory, and proposes a method to enforce consistency between density maps and object count for better performance. Through experiments, the effectiveness of the proposed weakly-supervised method is validated and shown to outperform existing solutions.
PATTERN RECOGNITION
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Qingyu Song et al.
Summary: This paper introduces a purely point-based framework for joint crowd counting and individual localization, with a new performance evaluation metric. By directly predicting point proposals with Point to Point Network (P2PNet), which is consistent with human annotations, it achieves significant improvement over existing methods.
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Zhihao Liu et al.
Summary: The research in crowd counting has advanced rapidly with the development of deep learning. To meet the high quality and quantity requirements in crowd counting research, researchers have organized the Vision Meets Drone Crowd Counting Challenge using a drone-captured dataset. This challenge has attracted many participants and paved the way for accelerating milestones in crowd counting.
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS (ICCVW 2021)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Yifan Yang et al.
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2020)
Article
Computer Science, Artificial Intelligence
Xialei Liu et al.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2019)
Proceedings Paper
Computer Science, Artificial Intelligence
Zan Shen et al.
2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2018)
Proceedings Paper
Computer Science, Artificial Intelligence
Elad Walach et al.
COMPUTER VISION - ECCV 2016, PT II
(2016)
Article
Automation & Control Systems
Min Fu et al.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2015)
Proceedings Paper
Computer Science, Artificial Intelligence
Ross Girshick
2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV)
(2015)
Article
Computer Science, Artificial Intelligence
Bo Wu et al.
INTERNATIONAL JOURNAL OF COMPUTER VISION
(2007)