We introduce blueprint separable convolutions (BSConv) as highly efficient building blocks for CNNs. They are motivated by quantitative analyses of kernel properties from trained models, which show the dominance of correlations along the depth axis. Based on our findings, we formulate a theoretical foundation from which we derive efficient implementations using only standard layers. Moreover, our approach provides a thorough theoretical derivation, interpretation, and justification for the application of depthwise separable convolutions (DSCs) in general, which have become the basis of many modern network architectures. Ultimately, we reveal that DSC-based architectures such as MobileNets implicitly rely on cross-kernel correlations, while our BSConv formulation is based on intra-kernel correlations and thus allows for a more efficient separation of regular convolutions. Extensive experiments on large-scale and fine-grained classification datasets show that BSConvs clearly and consistently improve MobileNets and other DSC-based architectures without introducing any further complexity. For fine-grained datasets, we achieve an improvement of up to 13.7 percentage points. In addition, if used as drop-in replacement for standard architectures such as ResNets, BSConv variants also outperform their vanilla counterparts by up to 9.5 percentage points on ImageNet. Code and models are available under https://github.com/zeiss-microscopy/BSConv.
翻译:我们引入了作为CNN高度高效构件的蓝图相分离的共变(BSConv),作为CNN的高效构件,其动机是从经过培训的模型对内核属性进行定量分析,这些模型显示在深度轴心上的相关性占主导地位。根据我们的调查结果,我们制定了一个理论基础,以便仅使用标准层来有效地实施。此外,我们的方法为应用深度分离的共变(DSCs Convolutions)提供了彻底的理论推理、解释和理由,这种应用已成为许多现代网络结构的基础。最终,我们透露,基于DSC的架构,例如移动网络,暗含地依赖跨内核关系,而我们的BSConv的配方则以内部关联为基础,从而能够更有效地分离经常性的共变相。关于大规模和精细的分类数据集的广泛实验表明,BSConvs明确和持续地改进移动网络和其他基于DSC的架构,而没有引入任何进一步的复杂性。关于精细的数据集,我们改进了以跨内核网络为主的关联关系,而我们的BSConvreal Pro replical im replus reforma 也用于了目前的版本。