AdaSemiCD: An Adaptive Semi-Supervised Change Detection Method Based on Pseudo-Label Evaluation

Change detection (CD) is an essential field in remote sensing, with a primary focus on identifying areas of change in bitemporal image pairs captured at varying intervals of the same region. The data annotation process for CD tasks is both time-consuming and labor-intensive. To better utilize the sc...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on geoscience and remote sensing Vol. 63; pp. 1 - 14
Main Authors: Ran, Lingyan, Wen, Dongcheng, Zhuo, Tao, Zhang, Shizhou, Zhang, Xiuwei, Zhang, Yanning
Format: Journal Article
Language:English
Published: New York IEEE 2025
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Adaptation models
Adaptive Learning
Annotations
Change detection
change detection (CD)
Data models
Entropy (Information theory)
Feature extraction
Filtering
Generative adversarial networks
Labels
mean teacher
Modules
Noise
Perturbation methods
pseudo-label
Remote sensing
Semi-supervised learning
semi-supervised learning (SSL)
Semisupervised learning
Training
ISSN:0196-2892, 1558-0644
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Change detection (CD) is an essential field in remote sensing, with a primary focus on identifying areas of change in bitemporal image pairs captured at varying intervals of the same region. The data annotation process for CD tasks is both time-consuming and labor-intensive. To better utilize the scarce labeled data and abundant unlabeled data, we introduce an adaptive semi-supervised learning (SSL) method, AdaSemiCD, to improve pseudo-label usage and optimize the training process. Initially, due to the extreme class imbalance inherent in CD, the model is more inclined to focus on the background class, and it is easy to confuse the boundary of the target object. Considering these two points, we develop a measurable evaluation metric for pseudo-labels that enhances the representation of information entropy by class rebalancing and amplification of ambiguous areas, assigning greater weights to prospective change objects. Subsequently, to enhance the reliability of sample wise pseudo-labels, we introduce the AdaFusion module, to dynamically identify the most uncertain region and substitute it with more trustworthy content. Lastly, to ensure better training stability, we introduce the AdaEMA module, which updates the teacher model using only batches of trusted samples. Experimental results on ten public CD datasets validate the efficacy and generalizability of our proposed adaptive training framework.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0196-2892
1558-0644
DOI:10.1109/TGRS.2025.3551504