A Semisupervised Arbitrary-Oriented SAR Ship Detection Network Based on Interference Consistency Learning and Pseudolabel Calibration
Yue Zhou, Xue Jiang, Zeming Chen, Lin Chen, Xingzhao Liu
Abstract
The rapid development of deep learning cannot be achieved without the support of abundant labeled data. However, obtaining such a large amount of annotated data needs the support of professionals in the field of synthetic aperture radar (SAR) image understanding, which leads to the scarcity of SAR datasets with annotations. The scarcity of annotations poses a bottleneck in the performance of SAR ship detectors based on deep learning. Recently, semi-supervised learning has become a hot paradigm, which can mine effective information from unlabeled data to further improve the performance of SAR ship detectors. However, existing semi-supervised SAR ship detection studies all adopted multi-stage semi-supervised frameworks, which are complex and inefficient. In this article, we first design an end-to-end semi-supervised framework for SAR ship detection. To overcome the strong interferences resulting from the imaging or quantization processes in SAR, we Introduce the interference consistency learning mechanism to enhance the model's robustness. To solve the complex background in the inshore scenario, a pseudo-label calibration network is designed to calibrate the pseudo-label according to the context knowledge around the ships. Based on the HRSID and the other four datasets, the superiority of the proposed approach over several state-of-the-art semi-supervised frameworks has been evaluated under various labeling ratios, i.e., 1%, 5%, 10%, and 100%.