Target scale information detection based on improved Faster R-CNN

Abstract

In order to obtain the scale information of ship targets effectively, we proposed an improved Faster R-CNN algorithm which integrated multi-scale region proposal and pooling of ROI, visual attention mechanism and rotation region regression and suppression, and ship targets can be positioned by the rotate quadrangle bounding boxes to obtain the scale information of them. Our improved model is based on the standard Faster R-CNN and is maintained through end-to-end training.