Abstract
Change detection (CD) using deep learning techniques is a prominent topic in the field of remote sensing (RS). However, existing methods require large amounts of labeled samples for supervised learning, which is time-consuming and labor-intensive. To address this challenge, semi-supervised learning (SSL) methods that utilize a limited number of labeled samples along with a large pool of unlabeled samples have emerged as a compelling solution. We propose a novel semi-supervised CD (SSCD) network which combines self-training and consistency regularization, namely STCRNet. During the self-training phase, STCRNet selects unlabeled samples with reliable pseudo-labels based on their prediction stability across different training epochs and the consistency between class activation maps (CAMs) and prediction results within the model. Then, we apply data augmentation to the reliable samples and enforce consistency regularization on the augmented samples using the pseudo-labels to enhance the network's robustness. Moreover, feature consistency regularization is applied to the remaining unlabeled samples with image perturbations, thereby broadening the feature space and improving the model's generalization performance. Experimental results on two widely-used datasets demonstrate that STCRNet achieves state-of-the-art (SOTA) performance, especially with a significantly small amount (5%∼10%) of labeled samples. STCRNet presents a promising solution for SSCD. The demo code is available at <uri>https://github.com/WangLukang/STCRNet</uri>.
Original language | English |
---|---|
Pages (from-to) | 1-12 |
Number of pages | 12 |
Journal | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
DOIs | |
Publication status | Accepted/In press - 2023 |
Keywords
- change detection (CD)
- consistency regularization
- Data models
- deep learning
- high-resolution images
- Perturbation methods
- Predictive models
- Reliability
- remote sensing
- self-training
- semi-supervised learning (SSL)
- Stability analysis
- Task analysis
- Training
ASJC Scopus subject areas
- Computers in Earth Sciences
- Atmospheric Science