基于 mask_rcnn 的舰船检测

最近想的idea 就发现有文章出来了，这世界真大，多多拜读吧
均为水文，大牛请忽略

paper1 INSHORE SHIP DETECTION BASED ON MASK R-CNN Shanlan

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8519123

1 、在maskrcnn的基础上加入了 soft-nms
2、战船与商船均能检测
3、基于mask 检测，可获得mask

论文中 说出自己改进的优势 和 和maskrcnn baseline 做对比

paper2 Attention-Based Convolutional Networks for Ship Detection in High-Resolution Remote Sensing Images

https://link.springer.com/chapter/10.1007/978-3-030-03341-5_31

Environmental information, like sea-land distribution, plays an important role in detecting ships from remote sensing images. However, the huge scale difference between environments and ship targets makes current CNN-based detection models hard to learn large-scale geographical information and focus on small targets at the same time. We propose an attention-based method by adding a Fully Convolutional Networks (FCN) to a detection networks as an attention branch to extract environmental features. Within a detection phase, the target detection branch is guided by the attention branch so as to focus on the potential target locations while in a training phase, the losses of other locations are simply ignored. We test our method on a public available remote sensing target detection dataset: LEVIR. By taking the classical Single Shot MultiBox Detector (SSD) as baseline, our method improves its detection accuracy in ship detection task while with an acceptable computational overhead.

提出了基于注意力机制，通过添加一个fcn 分支到检测网络去提取环境特征

paper2 Accurate Detection of Berthing Ship Target Based on Mask R-CNN 会议 忽略

https://www.spiedigitallibrary.org/conference-proceedings-of-spie/10836/1083602/Accurate-detection-of-berthing-ship-target-based-on-mask-R/10.1117/12.2326820.full?SSO=1

We adjusted the configuration of the region proposal network including the size of anchor frame and the threshold of non-maximum suppression according to the target morphology, so that the network generates a more accurate region of interest. Finally, the confidence levels, bounding-boxes and image masks of multi-objective generated concurrently. We performed experiments on self-made data sets which labeled from NWPU VHR-10 and produced good results, which proved the feasibility of this method in target detection of berthing ship target

调整了rpn anchor size的 大小和 nms的阈值等，其实就是改了该参数，这一篇