Neighborhood Feature Pooling for Remote Sensing Image Classification
This addresses remote sensing image classification for applications like land use mapping, though it appears incremental as a texture feature enhancement.
The paper tackles remote sensing image classification by proposing neighborhood feature pooling (NFP), a novel texture feature extraction method that aggregates local similarities, and results show it consistently improves performance across diverse datasets and architectures with minimal parameter overhead.
In this work, we propose neighborhood feature pooling (NFP) as a novel texture feature extraction method for remote sensing image classification. The NFP layer captures relationships between neighboring inputs and efficiently aggregates local similarities across feature dimensions. Implemented using convolutional layers, NFP can be seamlessly integrated into any network. Results comparing the baseline models and the NFP method indicate that NFP consistently improves performance across diverse datasets and architectures while maintaining minimal parameter overhead.