MDFN: Multi-scale deep feature learning network for object detection

作者：

Highlights：

• The paper proposes a new model that focuses on learning the deep features produced in the latter part of the network.

• Accurate detection results are achieved by making full use of the semantic and contextual information expressed by deep features.

• The proposed deep feature learning inception modules activate multi-scale receptive fields within a wide range at a single layer level.

• The paper demonstrates that features produced in the deeper part of networks have a prevailing impact on the accuracy of object detection.

摘要

•The paper proposes a new model that focuses on learning the deep features produced in the latter part of the network.•Accurate detection results are achieved by making full use of the semantic and contextual information expressed by deep features.•The proposed deep feature learning inception modules activate multi-scale receptive fields within a wide range at a single layer level.•The paper demonstrates that features produced in the deeper part of networks have a prevailing impact on the accuracy of object detection.

论文关键词：Deep feature learning,Multi-scale,Semantic and contextual information,Small and occluded objects

论文评审过程：Received 20 February 2019, Revised 5 September 2019, Accepted 4 December 2019, Available online 14 December 2019, Version of Record 29 December 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.107149