Deep-learning-based solution for data deficient satellite image segmentation

作者：

Highlights：

• Lower the impact of small labelled datasets on deep-learning-based methods.

• Proposed the Xception-based model using transfer learning.

• Proposed the novel steerable-filter-based model for handling rotation.

• Verified to achieve superior performance on small labelled dataset.

摘要

•Lower the impact of small labelled datasets on deep-learning-based methods.•Proposed the Xception-based model using transfer learning.•Proposed the novel steerable-filter-based model for handling rotation.•Verified to achieve superior performance on small labelled dataset.

论文关键词：Satellite image,Segmentation,Deep learning,Neural networks,Transfer learning

论文评审过程：Received 16 August 2020, Revised 27 July 2021, Accepted 6 November 2021, Available online 28 November 2021, Version of Record 13 December 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.116210