Drop an Octave: Reducing Spatial Redundancy in Convolutional Neural Networks with Octave Convolution
This addresses efficiency and accuracy challenges in image and video recognition for researchers and practitioners, offering a plug-and-play solution that is incremental but orthogonal to existing methods.
The paper tackles the problem of spatial redundancy in convolutional neural networks by introducing Octave Convolution (OctConv), which factorizes feature maps by frequency to process lower-frequency information at reduced resolution, resulting in improved accuracy and reduced computational cost, such as achieving 82.9% top-1 accuracy on ImageNet with 22.2 GFLOPs using a ResNet-152.
In natural images, information is conveyed at different frequencies where higher frequencies are usually encoded with fine details and lower frequencies are usually encoded with global structures. Similarly, the output feature maps of a convolution layer can also be seen as a mixture of information at different frequencies. In this work, we propose to factorize the mixed feature maps by their frequencies, and design a novel Octave Convolution (OctConv) operation to store and process feature maps that vary spatially "slower" at a lower spatial resolution reducing both memory and computation cost. Unlike existing multi-scale methods, OctConv is formulated as a single, generic, plug-and-play convolutional unit that can be used as a direct replacement of (vanilla) convolutions without any adjustments in the network architecture. It is also orthogonal and complementary to methods that suggest better topologies or reduce channel-wise redundancy like group or depth-wise convolutions. We experimentally show that by simply replacing convolutions with OctConv, we can consistently boost accuracy for both image and video recognition tasks, while reducing memory and computational cost. An OctConv-equipped ResNet-152 can achieve 82.9% top-1 classification accuracy on ImageNet with merely 22.2 GFLOPs.