FCN的分割效果看起来还不错,在深度学习之前,任意目标的图像分割还是一个比较棘手的问题,传统的图像算法比如满水填充water fill, 区域生长 只能适应背景简单的图像。
https://blog.csdn.net/weixin_34087301/article/details/87130567
C++调用mask rcnn进行实时检测--opencv4.0
https://blog.csdn.net/qq_29462849/article/details/85272575
OpenCV--FCN实现图像切割
https://blog.csdn.net/haiqiang1995/article/details/90114263
https://blog.csdn.net/daker_huang/article/details/87641756
Fully Convolutional Networks for Semantic Segmentation (CVPR 2015)
https://blog.csdn.net/tangwei2014/article/details/46882257
ENet(python+opencv)
https://www.jianshu.com/p/00215e4ceef7
代码和模型下载
https://github.com/e-lab/ENet-training
https://github.com/e-lab/ENet-training/releases/tag/v1.cs
实现
fcn8s没有天空分割
https://blog.csdn.net/wanggao_1990/article/details/87166102
代码
https://blog.csdn.net/a_l_a_n/article/details/86424175
ENet论文(CVPR2016)
ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation
ENet是目前实时性最快的semantic segmentation
https://zhuanlan.zhihu.com/p/31379024
深度学习分割网络Unet
https://blog.csdn.net/Formlsl/article/details/80373200