Global and Local Contrastive Self-Supervised Learning for Semantic Segmentation of HR Remote Sensing Images

• 论文名称: Global and Local Contrastive Self-Supervised Learning for Semantic Segmentation of HR Remote Sensing Images
• Link: https://arxiv.org/abs/2106.10605
• Published in: IEEE Transactions on Geoscience and Remote Sensing 2022
• Type: Pretrain
• Code/Project: https://github.com/GeoX-Lab/G-RSIM
• 备注: 自监督，对比学习
• Backbone: CNN
• Backbone 1: CNN
• 下游任务: Semantic Segmentation
• 下游任务 1: Semantic Segmentation
• Short Summary: 基于对比学习对于像素级别的任务适配不好，改论文提出了一种基于全局和局部对比学习的方法，以更好的适应像素级任务