Towards Better Exploiting Convolutional Neural Networks for Remote Sensing Scene Classification

TL;DR

This paper analyzes how to best utilize pre-trained convolutional neural networks for remote sensing scene classification, finding fine-tuning with linear SVM yields the best results and achieves state-of-the-art performance.

Contribution

It systematically compares three strategies for exploiting ConvNets in remote sensing, highlighting fine-tuning as the most effective approach.

Findings

Fine-tuning outperforms other strategies.

Features from fine-tuned ConvNets with linear SVM give top results.

Achieved state-of-the-art accuracy on three remote sensing datasets.

Abstract

We present an analysis of three possible strategies for exploiting the power of existing convolutional neural networks (ConvNets) in different scenarios from the ones they were trained: full training, fine tuning, and using ConvNets as feature extractors. In many applications, especially including remote sensing, it is not feasible to fully design and train a new ConvNet, as this usually requires a considerable amount of labeled data and demands high computational costs. Therefore, it is important to understand how to obtain the best profit from existing ConvNets. We perform experiments with six popular ConvNets using three remote sensing datasets. We also compare ConvNets in each strategy with existing descriptors and with state-of-the-art baselines. Results point that fine tuning tends to be the best performing strategy. In fact, using the features from the fine-tuned ConvNet with linear SVM obtains the best results. We also achieved state-of-the-art results for the three datasets used.