Abstract
In remote-sensing image (RSI) semantic segmentation, the dependence on large-scale and pixel-level annotated data has been a critical factor restricting its development. In this letter, we propose an unsupervised semantic segmentation network embedded with geometry consistency (UGCNet) for RSIs, which imports the adversarial-generative learning strategy into a semantic segmentation network. The proposed UGCNet can be trained on a source-domain dataset and achieve accurate segmentation results on a different target-domain dataset. Furthermore, for refining the remote-sensing target geometric representation such as densely distributed buildings, we propose a geometry-consistency (GC) constraint that can be embedded in both image-domain adaptation process and semantic segmentation network. Therefore, our model could achieve cross-domain semantic segmentation with target geometric property preservation. The experimental results on Massachusetts and Inria buildings datasets prove that the proposed unsupervised UGCNet could achieve a very comparable segmentation accuracy with the fully supervised model, which validates the effectiveness of the proposed method.
Highlights
S EMANTIC segmentation aims to assign a label to every single pixel in the image
In this letter we propose an unsupervised semantic segmentation network (UGCNet) for remote sensing image (RSI) that preserves image geometric structures during the adaptation process
Our contributions are summarized as follows: (1) Through the adaptation between source-domain images and target-domain images, we propose a novel unsupervised semantic segmentation method (UGCNet) for remote sensing images
Summary
S EMANTIC segmentation aims to assign a label to every single pixel in the image. Due to the rapid development of deep learning in recent years, numerous semantic segmentation algorithms like [1], [2] for RSIs are proposed. One rational strategy for unsupervised semantic segmentation task is to utilize image domain adaptation. In this letter we propose an unsupervised semantic segmentation network (UGCNet) for RSIs that preserves image geometric structures during the adaptation process. In the GSN, a pixel-level segmentation network is trained on the transferred sourcedomain images, and its model can be applied on the targetdomain images. (1) Through the adaptation between source-domain images and target-domain images, we propose a novel unsupervised semantic segmentation method (UGCNet) for remote sensing images. (2) We propose a geometry-consistency constraint that could be embedded in both image translation and semantic segmentation networks, which simultaneously preserves semantic structures of source-domain images and improves semantic segmentation performance on target domain.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
