Zhang Yi, Song Wei, Shao Mingyue, Liu Xiangchun
School of Information and Engineering, Minzu University of China, Beijing 100081, China.
Language Information Security Research Center, Institute of National Security MUC, Minzu University of China, Beijing 100081, China.
Sensors (Basel). 2024 Sep 14;24(18):5974. doi: 10.3390/s24185974.
This study focuses on the problem of dense object counting. In dense scenes, variations in object scales and uneven distributions greatly hinder counting accuracy. The current methods, whether CNNs with fixed convolutional kernel sizes or Transformers with fixed attention sizes, struggle to handle such variability effectively. Lower-resolution features are more sensitive to larger objects closer to the camera, while higher-resolution features are more efficient for smaller objects further away. Thus, preserving features that carry the most relevant information at each scale is crucial for improving counting precision. Motivated by this, we propose a multi-resolution scale feature fusion-based universal density counting network (MRSNet). It utilizes independent modules to process high- and low-resolution features, adaptively adjusts receptive field sizes, and incorporates dynamic sparse attention mechanisms to optimize feature information at each resolution, by integrating optimal features across multiple scales into density maps for counting evaluation. Our proposed network effectively mitigates issues caused by large variations in object scales, thereby enhancing counting accuracy. Furthermore, extensive quantitative analyses on six public datasets demonstrate the algorithm's strong generalization ability in handling diverse object scale variations.
本研究聚焦于密集物体计数问题。在密集场景中，物体尺度的变化和分布不均极大地阻碍了计数精度。当前的方法，无论是具有固定卷积核大小的卷积神经网络（CNNs）还是具有固定注意力大小的Transformer，都难以有效处理这种变化性。较低分辨率的特征对靠近相机的较大物体更敏感，而较高分辨率的特征对较远的较小物体更有效。因此，在每个尺度上保留携带最相关信息的特征对于提高计数精度至关重要。受此启发，我们提出了一种基于多分辨率尺度特征融合的通用密度计数网络（MRSNet）。它利用独立模块处理高分辨率和低分辨率特征，自适应调整感受野大小，并结合动态稀疏注意力机制在每个分辨率下优化特征信息，通过将多个尺度的最优特征整合到密度图中进行计数评估。我们提出的网络有效地减轻了由物体尺度的巨大变化引起的问题，从而提高了计数精度。此外，在六个公共数据集上进行的广泛定量分析证明了该算法在处理各种物体尺度变化方面具有很强的泛化能力。
