SCNet论文详解:Improving Convolutional Networks with Self-calibrated Convolutions
《ImprovingConvolutionalNetworkswithSelf-calibratedConvolutions》是2020年CVPR的论文,作者来自于南开大学程明明团队。最近各种卷积注意力组合的模块工作层出不穷,性能涨点明显,包括之前的Res2Net、李沐团队的ResNeSt,应该是近期的热点方向。论文地址:http://mftp.mmcheng.net/Papers/20cvprS