Multi-scale features fused network with multi-level supervised path for crowd counting

作者：

Highlights：

• Multi-level dilated convolution module supervises the whole network at multi-level.

• Soft spatial-channel attention module produces a saliency weight map of same size.

• Our method achieves better experimental results on four challenging datasets.

摘要

•Multi-level dilated convolution module supervises the whole network at multi-level.•Soft spatial-channel attention module produces a saliency weight map of same size.•Our method achieves better experimental results on four challenging datasets.

论文关键词：Crowd counting,Multi-level supervision,Soft spatial-channel attention module,Multi-level dilated convolution module

论文评审过程：Received 7 August 2020, Revised 8 March 2022, Accepted 18 March 2022, Available online 29 March 2022, Version of Record 5 April 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.116949