Convolution of Convolution: Let Kernels Spatially Collaborate
In the biological visual pathway, especially the retina, neurons are tiled along spatial dimensions with the electrical coupling as their local association, while in a convolution layer, kernels are placed along the channel dimension singly. We propose Convolution of Convolution, associating kernels in a layer and letting them collaborate spatially. With this method, a layer can provide feature maps with extra transformations and learn its kernels together instead of isolatedly. It is only used during training, bringing in negligible extra costs; and can be re-parameterized to common convolution before testing, boosting performance gratuitously in tasks like classification, detection and segmentation. Our method works even better when large receptive fields are demanded. The code is available on site: https://github.com/Genera1Z/ConvolutionOfConvolution.
Code (1)
Methods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Fast and High-Quality Image Denoising via Malleable Convolutions
Most image denoising networks apply a single set of static convolutional kernels across the entire input image. This is sub-optimal for natural images, as they often consist of heterogeneous visual patterns. Dynamic conv…
DenoisingImage DenoisingImage RestorationVocal Bursts Intensity PredictionA Constrained Deformable Convolutional Network for Efficient Single Image Dynamic Scene Blind Deblurring with Spatially-Variant Motion Blur Kernels Estimation
Most existing deep-learning-based single image dynamic scene blind deblurring (SIDSBD) methods usually design deep networks to directly remove the spatially-variant motion blurs from one inputted motion blurred image, wi…
DeblurringImage RestorationSACB-Net: Spatial-awareness Convolutions for Medical Image Registration
Deep learning-based image registration methods have shown state-of-the-art performance and rapid inference speeds. Despite these advances, many existing approaches fall short in capturing spatially varying information in…
Image RegistrationMedical Image RegistrationDISK: Differentiable Sparse Kernel Complex for Efficient Spatially-Variant Convolution
Image convolution with complex kernels is a fundamental operation in photography, scientific imaging, and animation effects, yet direct dense convolution is computationally prohibitive on resource-limited devices. Existi…
Neural Field Convolutions by Repeated Differentiation
Neural fields are evolving towards a general-purpose continuous representation for visual computing. Yet, despite their numerous appealing properties, they are hardly amenable to signal processing. As a remedy, we presen…