• Home
  • Search
  • CDEST: Class Distinguishability-Enhanced Self-Training Method for Adopting Pre-Trained Models to Downstream Remote Sensing Image Semantic Segmentation
  • Cite Icon1
  • https://doi.org/10.3390/rs16071293Copy DOI Icon

CDEST: Class Distinguishability-Enhanced Self-Training Method for Adopting Pre-Trained Models to Downstream Remote Sensing Image Semantic Segmentation

Show More
  • Abstract
  • Highlights & Summary
  • PDF
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

The self-supervised learning (SSL) technique, driven by massive unlabeled data, is expected to be a promising solution for semantic segmentation of remote sensing images (RSIs) with limited labeled data, revolutionizing transfer learning. Traditional ‘local-to-local’ transfer from small, local datasets to another target dataset plays an ever-shrinking role due to RSIs’ diverse distribution shifts. Instead, SSL promotes a ‘global-to-local’ transfer paradigm, in which generalized models pre-trained on arbitrarily large unlabeled datasets are fine-tuned to the target dataset to overcome data distribution shifts. However, the SSL pre-trained models may contain both useful and useless features for the downstream semantic segmentation task, due to the gap between the SSL tasks and the downstream task. To adapt such pre-trained models to semantic segmentation tasks, traditional supervised fine-tuning methods that use only a small number of labeled samples may drop out useful features due to overfitting. The main reason behind this is that supervised fine-tuning aims to map a few training samples from the high-dimensional, sparse image space to the low-dimensional, compact semantic space defined by the downstream labels, resulting in a degradation of the distinguishability. To address the above issues, we propose a class distinguishability-enhanced self-training (CDEST) method to support global-to-local transfer. First, the self-training module in CDEST introduces a semi-supervised learning mechanism to fully utilize the large amount of unlabeled data in the downstream task to increase the size and diversity of the training data, thus alleviating the problem of biased overfitting of the model. Second, the supervised and semi-supervised contrastive learning modules of CDEST can explicitly enhance the class distinguishability of features, helping to preserve the useful features learned from pre-training while adapting to downstream tasks. We evaluate the proposed CDEST method on four RSI semantic segmentation datasets, and our method achieves optimal experimental results on all four datasets compared to supervised fine-tuning as well as three semi-supervised fine-tuning methods.

Loading PDF

Similar Papers
  • PDF
  • Research Article
  • Citations16

A Generic Self-Supervised Learning (SSL) Framework for Representation Learning from Spectral–Spatial Features of Unlabeled Remote Sensing Imagery

  • Nov 03, 2023
  • Remote Sensing
  • Xin Zhang +1
  • Research Article
  • Citations174

Global and Local Contrastive Self-Supervised Learning for Semantic Segmentation of HR Remote Sensing Images

  • Jan 01, 2022
  • IEEE Transactions on Geoscience and Remote Sensing
  • Haifeng Li +6
  • Research Article
  • Citations94

CMID: A Unified Self-Supervised Learning Framework for Remote Sensing Image Understanding

  • Jan 01, 2023
  • IEEE Transactions on Geoscience and Remote Sensing
  • Dilxat Muhtar +4
  • PDF
  • Research Article
  • Citations45

Deep Learning Methods for Semantic Segmentation in Remote Sensing with Small Data: A Survey

  • Oct 16, 2023
  • Remote Sensing
  • Anzhu Yu +9
  • Preprint Article

Generalist Geospatial Foundation Model PhilEO on Satellite Sentinel-2 MajorTOM Multi-Spectral Data

  • Mar 15, 2025
  • Nikolaos Dionelis +9
  • PDF
  • Research Article
  • Citations4

Deep learning based on self-supervised pre-training: Application on sandstone content prediction

  • Jan 10, 2023
  • Frontiers in Earth Science
  • Chong Ming Wang +5
  • Research Article

Spatial-Frequency Contrastive Learning for Self-Supervised Semantic Segmentation of Remote Sensing Images

  • Jan 01, 2026
  • IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
  • Yujia Fu +3
  • Conference Article
  • Citations9

DATA: Domain-Aware and Task-Aware Self-supervised Learning

  • Jun 01, 2022
  • Qing Chang +6
  • Research Article
  • Citations1

Contrastive Masked Feature Modeling for Self-Supervised Representation Learning of High-Resolution Remote Sensing Images

  • Feb 17, 2026
  • Remote Sensing
  • Shiyan Pang +4
  • Research Article
  • Citations286

RingMo: A Remote Sensing Foundation Model With Masked Image Modeling

  • Jan 01, 2023
  • IEEE Transactions on Geoscience and Remote Sensing
  • Xian Sun +14
  • Conference Article
  • Citations12

Conformer-Based Self-Supervised Learning For Non-Speech Audio Tasks

  • May 23, 2022
  • Sangeeta Srivastava +6
  • Conference Article
  • Citations2

A Novel Multi-Task Self-Supervised Representation Learning Paradigm

  • May 28, 2021
  • Control theory & applications
  • Yinggang Li +5
  • Research Article
  • Citations11

Joint Self-Supervised Image-Volume Representation Learning with Intra-inter Contrastive Clustering

  • Jun 26, 2023
  • Proceedings of the AAAI Conference on Artificial Intelligence
  • Duy M H Nguyen +9
  • Supplementary Content
  • Citations259

Self-supervised Learning: A Succinct Review

  • Jan 01, 2023
  • Archives of Computational Methods in Engineering
  • Veenu Rani +4
  • PDF
  • Research Article
  • Citations20

Self-supervised learning for remote sensing scene classification under the few shot scenario

  • Jan 09, 2023
  • Scientific Reports
  • Najd Alosaimi +4
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.