Abstract: We propose reparameterized refocusing convolution (RefConv) as a replacement for regular convolutional layers, which is a plug-and-play module to improve the performance without any ...
Offical implementation of "Strip R-CNN: Large Strip Convolution for Remote Sensing Object Detection" we also add our config in https://github.com/zcablii/LSKNet and ...
Abstract: Pixel-level adaptive convolution, which overcomes the deficiency of the spatial-invariance of standard convolution, is always limited to performing feature extraction from local patches and ...