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

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

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 datas...

Full description

Bibliographic Details
Main Authors: Ming Zhang, Xin Gu, Ji Qi, Zhenshi Zhang, Hemeng Yang, Jun Xu, Chengli Peng, Haifeng Li
Format: Article
Language:English
Published: MDPI AG 2024-04-01
Series:Remote Sensing
Subjects:
semantic segmentation
remote sensing (RS)
transfer learning
fine-tuning method
contrastive learning
self-training
Online Access:https://www.mdpi.com/2072-4292/16/7/1293

Internet

https://www.mdpi.com/2072-4292/16/7/1293

Similar Items