Convolutional neural networks for multispectral pedestrian detection

作者：

Highlights：

• We designed a network that integrates two sub-networks to handle different scales of pedestrians.

• We exploit the NIN fusion strategy in our network model and propose a novel fusion R-FCN networks.

• Our model surpasses the latest method on KAIST multispectral pedestrian dataset (KAIST).

摘要

•We designed a network that integrates two sub-networks to handle different scales of pedestrians.•We exploit the NIN fusion strategy in our network model and propose a novel fusion R-FCN networks.•Our model surpasses the latest method on KAIST multispectral pedestrian dataset (KAIST).

论文关键词：Multispectral pedestrian detection,R-FCN,Network-in-network

论文评审过程：Received 15 February 2019, Revised 24 December 2019, Accepted 24 December 2019, Available online 7 January 2020, Version of Record 7 January 2020.

论文官网地址：https://doi.org/10.1016/j.image.2019.115764