Dynamic and Adaptive Self-Training for Semi-Supervised Remote Sensing Image Semantic Segmentation
Jidong Jin, Wanxuan Lu, Hongfeng Yu, Xuee Rong, Xian Sun, Yirong Wu
Abstract
Remote sensing technology has made remarkable progress, providing a wealth of data for various applications, such as ecological conservation and urban planning. However, the meticulous annotation of this data is labor-intensive, leading to a shortage of labeled data, particularly in tasks like semantic segmentation. Semi-supervised methods, combining consistency regularization with self-training, offer a solution to efficiently utilize labeled and unlabeled data. However, these methods encounter challenges due to imbalanced data ratios. To tackle these challenges, we introduce a self-training approach named <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DAST</i> ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">D</i> ynamic and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">A</i> daptive <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">S</i> elf- <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</i> raining), which is combined with dynamic pseudo-label sampling, distribution matching, and adaptive threshold updating. Dynamic pseudo-label sampling is tailored to address the issue of class distribution imbalance by giving priority to classes with fewer samples. Meanwhile, distribution matching and adaptive threshold updating aim to reduce distribution disparities by adjusting model predictions across augmented images within the framework of consistency regularization, ensuring they align with the actual data distribution. Experiment results on the Potsdam and iSAID datasets demonstrate that DAST effectively balances class distribution, aligns model predictions with data distribution, and stabilizes pseudo-labels, leading to state-of-the-art performance on both datasets. These findings highlight the potential of DAST in overcoming the challenges associated with significant disparities in labeled-to-unlabeled data ratios.