We propose a new convolution called Dynamic Region-Aware Convolution (DRConv), which can automatically assign multiple filters to corresponding spatial regions where features have similar representation. In this way, DRConv outperforms standard convolution in modeling semantic variations. Standard convolutional layer can increase the number of filers to extract more visual elements but results in high computational cost. More gracefully, our DRConv transfers the increasing channel-wise filters to spatial dimension with learnable instructor, which not only improve representation ability of convolution, but also maintains computational cost and the translation-invariance as standard convolution dose. DRConv is an effective and elegant method for handling complex and variable spatial information distribution. It can substitute standard convolution in any existing networks for its plug-and-play property, especially to power convolution layers in efficient networks. We evaluate DRConv on a wide range of models (MobileNet series, ShuffleNetV2, etc.) and tasks (Classification, Face Recognition, Detection and Segmentation). On ImageNet classification, DRConv-based ShuffleNetV2-0.5x achieves state-of-the-art performance of 67.1% at 46M multiply-adds level with 6.3% relative improvement.
翻译:我们提出一个新的革命,称为动态区域-磁器革命(DRConv),它可以自动地将多个过滤器分配给具有类似特征的相应的空间区域。 这样, DRConv在建模语义变异中优于标准演进。 标准革命层可以增加文件员的数量, 提取更多的视觉元素, 但却导致高计算成本。 更优雅地, 我们的DRConv将不断增长的带频道的过滤器转移到空间层面, 由可学习的教员提供, 这不仅可以提高演化的代表性, 还可以维持计算成本和转换性能, 作为标准演动剂量。 DRConv是处理复杂和变异空间信息分布的有效和优雅的方法。 它可以取代任何现有网络的标准演动, 特别是用于高效网络中的动力变动层。 我们评估DRConvonv关于多种模型( 移动网络系列, ShuffleNetV2, 等) 和任务( 分类、 面识别、探测和分解) 。 关于图像网络分类, DRCR-Con- Churflele- weardleNet- 6.M.5xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx