<italic>DiverseNet:</italic> Decision Diversified Semi-Supervised Semantic Segmentation Networks for Remote Sensing Imagery
Semi-supervised learning (SSL) aims to help reduce the cost of the manual labeling process by leveraging a substantial pool of unlabeled data alongside a limited set of labeled data during the training phase. Since pixel-level manual labeling in large-scale remote sensing imagery is expensive and ti...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/11015984/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Semi-supervised learning (SSL) aims to help reduce the cost of the manual labeling process by leveraging a substantial pool of unlabeled data alongside a limited set of labeled data during the training phase. Since pixel-level manual labeling in large-scale remote sensing imagery is expensive and time-consuming, SSL has become a widely used solution to deal with this. However, the majority of existing SSL frameworks, especially various teacher–student frameworks, are too bulky to run efficiently on a GPU with limited memory. There is still a lack of lightweight SSL frameworks and efficient perturbation methods to promote the diversity of training samples and enhance the precision of pseudolabels during training. In order to fill this gap, we proposed a simple, lightweight, and efficient SSL architecture named <italic>DiverseHead</italic>, which promotes the utilisation of multiple decision heads instead of multiple whole networks. Another limitation of most existing SSL frameworks is the insufficient diversity of pseudolabels, as they rely on the same network architecture and fail to explore different structures for generating pseudolabels. To solve this issue, we propose <italic>DiverseModel</italic> to explore and analyze different networks in parallel for SSL to increase the diversity of pseudolabels. The two proposed methods, namely <italic>DiverseHead</italic> and <italic>DiverseModel</italic>, both achieve competitive semantic segmentation performance in four widely used remote sensing imagery datasets compared to state-of-the-art SSL methods. Meanwhile, the proposed lightweight DiverseHead architecture can be easily applied to various state-of-the-art SSL methods while further improving their performance. |
|---|---|
| ISSN: | 1939-1404 2151-1535 |