Abstract:As a content management technique, remote sensing (RS) scene classification (RSSC) always attracts researchers’ attention. In the past decades, many successful methods have been proposed. Nevertheless, their prerequisite is that there are large labeled datasets, which is a strict demand in practice. To resolve this contradiction, developing RSSC models with the help of few-shot learning (FSL) has become popular. Due to lacking prior knowledge, most of the existing few-shot RSSC models pay attention to the learning algorithm. However, they do not attach importance to the complex contents within RS scenes and the intricate interclass/intraclass relations between RS scenes. This would influence their performance negatively. In this article, we propose a new few-shot RSSC model named multipretext-task prototypes guided dynamic contrastive learning network (MPCL-Net). MPCL-Net consists of a multipretext tasks generation submodule, a deep feature learning submodule, and a joint optimization submodule. First, two RS-oriented pretext tasks are constructed under the self-supervised learning (SSL) framework in the multipretext tasks generation submodule, which aims to explore multiscale and rotation-invariant information from RS scenes. Second, a simple convolutional neural network (CNN) is developed in the deep feature learning submodule to transform the RS scenes into visual features. Third, three loss functions are formulated and integrated into the joint optimization submodule. Their goals are to fully capture the diverse land covers within RS scenes and compact/separate the intraclass/interclass samples with limited supervision. Finally, our MPCL-Net can be trained in a meta way. The positive results counted on the three public RS scene datasets confirm that our MPCL-Net is helpful to RSSC tasks under the few-shot scenario. Our source codes are available at https://github.com/TangXu-Group/Remote-Sensing-Images-Classification/tree/main/MPCL.

NQ-Protonet: Noisy Query Prototypical Network for Few-Shot Remote Sensing Scene Classification.

SPNet: Siamese-Prototype Network for Few-Shot Remote Sensing Image Scene Classification

Hierarchical Representation based Query-Specific Prototypical Network for Few-Shot Image Classification

MFGNet: Multibranch Feature Generation Networks for Few-Shot Remote Sensing Scene Classification.

ALPN: Active-Learning-Based Prototypical Network for Few-Shot Hyperspectral Imagery Classification

Prototype Calibration with Feature Generation for Few-Shot Remote Sensing Image Scene Classification

Heterogeneous Prototype Distillation With Support-Query Correlative Guidance for Few-Shot Remote Sensing Scene Classification

RS-SSKD: Self-Supervision Equipped with Knowledge Distillation for Few-Shot Remote Sensing Scene Classification

Few-Shot Scene Classification of Optical Remote Sensing Images Leveraging Calibrated Pretext Tasks

Prototype Rectification with Region-Wise Foreground Enhancement for Few-Shot Classification.

MKN: Metakernel Networks for Few Shot Remote Sensing Scene Classification

Query-guided Prototype Evolution Network for Few-Shot Segmentation

Query-guided Support Prototypes for Few-shot 3D Indoor Segmentation

SGMNet: Scene Graph Matching Network for Few-Shot Remote Sensing Scene Classification

TDNet: A Novel Transductive Learning Framework with Conditional Metric Embedding for Few-shot Remote Sensing Image Scene Classification

HiReNet: Hierarchical-Relation Network for Few-Shot Remote Sensing Image Scene Classification

RS-MetaNet: Deep meta metric learning for few-shot remote sensing scene classification

Not Just Learning from Others but Relying on Yourself: A New Perspective on Few-Shot Segmentation in Remote Sensing

PRSN: Prototype resynthesis network with cross-image semantic alignment for few-shot image classification

Multi-pretext-task Prototypes Guided Dynamic Contrastive Learning Network for Few-shot Remote Sensing Scene Classification

Prototype Refinement Network for Few-Shot Segmentation