Neighborhood Feature Pooling for Remote Sensing Image Classification

2025/10/29 by Nia, Fahimeh Orvati, Mohammadi, Amirmohammad, Kharsa, Salim Al +3
#68T07 #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #I.2.10 #I.4.8 #Image and Video Processing (eess.IV) #electronic engineering #information engineering

paper · doi:10.48550/arxiv.2510.25077

Abstract

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.

Citations

• Lacunarity Pooling Layers for Plant Image Classification using Texture Analysis
• Exploring the Sharpened Cosine Similarity
• RADAM: Texture Recognition through Randomized Aggregated Encoding of Deep Activation Maps
• A ConvNet for the 2020s
• A ConvNet for the 2020s
• Local Similarity Pattern and Cost Self-Reassembling for Deep Stereo Matching Networks
• ResNet strikes back: An improved training procedure in timm
• MSMatch: Semi-Supervised Multispectral Scene Classification with Few Labels
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• In-domain representation learning for remote sensing
• Deep Texture Manifold for Ground Terrain Recognition
• Squeeze-and-Excitation Networks
• EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification
• Deep TEN: Texture Encoding Network
• Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
• An open access repository of images on plant health to enable the development of mobile disease diagnostics
• Silhouettes: A graphical aid to the interpretation and validation of cluster analysis

Related

• DMFNet: Dual-Backbone Multiscale Fusion Network for Urban Scene Classification
• Convolution Neural Network Architecture Learning for Remote Sensing Scene Classification
• Effective Use of Dilated Convolutions for Segmenting Small Object Instances in Remote Sensing Imagery
• Leveraging feature communication in federated learning for remote sensing image classification
• Tiling and Stitching Segmentation Output for Remote Sensing: Basic Challenges and Recommendations