 This paper proposes a novel approach to improving the performance of instant segmentation and high-resolution remote sensing, HRS, images. Specifically, it modifies the structure of the one-stage fast detection network to better adapt to the task of ship target segmentation while also improving the efficiency of the network. Additionally, two feature optimization modules are added to the backbone network to further improve its feature learning capabilities. Furthermore, the network's feature fusion structure is modified to increase the prediction ability of multi-scale targets while reducing the amount of model calculation. Lastly, extensive validation experiments were conducted on the HRSID and SSDD datasets, demonstrating improved instance mask accuracy and enhanced segmentation efficiency. As such, this proposed model is a more accurate and efficient solution for instance segmentation in HRS imagery. This article was authored by Muhammad Yasser, Lili Jean, Shan Wei Lu, and others. We are article.tv, links in the description below.