Crowd Counting by Multi-Scale Dilated Convolution Networks

Article Zoology

Feasibility of using convolutional neural networks for individual-identification of wild Asian elephants

Elgiriyage M. K. de Silva et al.

Summary: Individual identification is crucial for research in behavior, ecology, and conservation. Deep-learning methods, particularly Convolutional Neural Networks, show promise in automating this task. Different CNN models were evaluated for identifying Asian elephants through photographs, with VGG16 performing best on body and face datasets, and Xception excelling on ear dataset.

MAMMALIAN BIOLOGY (2022)

添加到收藏夹

Article Engineering, Civil

Image-Based Crowd Stability Analysis Using Improved Multi-Column Convolutional Neural Network

Rongyong Zhao et al.

Summary: This paper proposes a novel model for crowd stability analysis using images from a real-time video surveillance system. By enhancing the accuracy of human head recognition for crowd counting and density estimation, a four-column convolutional neural network (4C-CNN) is developed. This model calculates crowd density in different areas with image rectification against perspective distortion, and utilizes a stability criterion for quantitative computation dynamically. Additionally, experiments demonstrate that this improved CNN model outperforms typical multi-column CNN models for crowd counting, and the effectiveness of the crowd stability analysis model is validated at Shanghai Hongqiao Railway Station.

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS (2022)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Boosting Crowd Counting via Multifaceted Attention

Hui Lin et al.

Summary: This paper presents a Multifaceted Attention Network (MAN) to deal with the challenging task of crowd counting. The proposed network incorporates global attention, learnable local attention, and instance attention to enhance the performance of Transformer models in local spatial relation encoding.

2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022) (2022)

添加到收藏夹

Article Multidisciplinary Sciences

Convolutional Neural Network for Crowd Counting on Metro Platforms

Jun Zhang et al.

Summary: This paper addresses the issue of passenger flow detection on metro platforms and proposes a CNN-based network to accurately count people. Through experiments on a self-built dataset and standard datasets, the proposed method shows good performance and minimal estimation error.

SYMMETRY-BASEL (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Rethinking Counting and Localization in Crowds: A Purely Point-Based Framework

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)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence