Abstract:High-resolution remote sensing image scene classification problem has attracted lots of attentions due to its crucial role in a wide range of applications. Recently, many convolutional neural network (CNN) based methods have significantly boosted the performance of image scene classification. However, most of these algorithms only make use of global features learned in the fully connected layers of a CNN, neglecting local features learned in the convolutional layers that is of crucial importance for some remote sensing scenes. Therefore, a sparse representation framework is proposed to make use of both local and global features learned in a CNN for classification of remote sensing scenes by balancing sparse representation based classifiers using these two kinds of features. Specially, in order to reduce redundancy of local features learned in a convolutional layer, the most effective local feature is generated from each convolutional layer using global average pooling and these selected local features of di fferent convolutional layers are cascaded to form local feature representation of the scene. Finally, experimental results on UC-Merced and WHU-RS19 datasets demonstrate that fusing global and local features in a CNN using the proposed sparse representation framework can certainly improve the performance of classification using single kind of features. Moreover, the proposed global average pooling strategy is very effective to fuse local features select most representative features from convolutional layers of a CNN.

Combining CNN with Hand-Crafted Features for Image Classification

Research on Image Classification Method of Features of Combinatorial Convolution

On Fusing the Latent Deep CNN Feature for Image Classification.

Handcrafted Local Features are Convolutional Neural Networks

Invariant Feature Extraction for Image Classification Via Multi-Channel Convolutional Neural Network

Multilevel Features Fusion In Deep Convolutional Neural Networks

Fusing Deep Local And Global Features For Remote Sensing Image Scene Classification

Combining transformers with CNN for multi-focus image fusion

A Late Fusion Approach for Harnessing Multi-Cnn Model High-Level Features

Convolutional Neural Network Image Classification Method Integrating Classic Image Features

A FEATURE EMBEDDING STRATEGY FOR HIGH-LEVEL CNN REPRESENTATIONS FROM MULTIPLE CONVNETS

Fusion of Dilated Convolution in CNN and Transformer Networks for Hyperspectral Image Classification

A Multi-Feature Fusion Model Based on Denoising Convolutional Neural Network and Attention Mechanism for Image Classification

Improving the Performance of Convolutional Neural Networks by Fusing Low-Level Features with Different Scales in the Preceding Stage

Feeding Hand-Crafted Features for Enhancing the Performance of Convolutional Neural Networks

Bridging CNN and Transformer With Cross-Attention Fusion Network for Hyperspectral Image Classification

Very High Resolution Images Classification by Fusing Deep Convolutional Neural Networks

High-Resolution Remote Sensing Image Retrieval Based on Classification-Similarity Networks and Double Fusion

Effect of fusing features from multiple DCNN architectures in image classification.

Improved Image Classification With Token Fusion

Automatic Image Annotation Via Combining Low-level Colour Feature with Features Learned from Convolutional Neural Networks