 The paper proposes a new method called dual polarization information, guided network, DPIG net, to improve the performance of SAR ship classification by taking advantage of both horizontal and vertical polarizations. Firstly, the authors developed a polarization channel cross-attention framework, PCCAF, to extract features from the two polarizations. Secondly, they introduced a novel dilated residual dense learning framework, DRDLF, to combine the extracted features from PCCAF and further improve the classification accuracy. Finally, the authors evaluated their approach on the open SAR ship dataset and found that DPIG net outperformed all other methods, achieving a higher accuracy than the baseline model. This article was authored by Zikang Xiao, Tianwen Zhang, and Xiao Kuo.