DO-Conv: Depthwise Over-Parameterized Convolutional Layer

Jinming Cao

doi:10.1109/TIP.2022.3175432

DO-Conv: Depthwise Over-Parameterized Convolutional Layer

2022-01-01

95

OA

AI

DOI：10.1109/TIP.2022.3175432

Original

Original request for help

Save

Abstract

En 中文

Convolutional layers are the core building blocks of Convolutional Neural Networks (CNNs). In this paper, we propose to augment a convolutional layer with an additional depthwise convolution, where each input channel is convolved with a different 2D kernel. The composition of the two convolutions constitutes an over-parameterization, since it adds learnable parameters, while the resulting linear operation can be expressed by a single convolution layer. We refer to this depthwise over-parameterized convolutional layer as DO-Conv, which is a novel way of over-parameterization. We show with extensive experiments that the mere replacement of conventional convolutional layers with DO-Conv layers boosts the performance of CNNs on many classical vision tasks, such as image classification, detection, and segmentation. Moreover, in the inference phase, the depthwise convolution is folded into the conventional convolution, reducing the computation to be exactly equivalent to that of a convolutional layer without over-parameterization. As DO-Conv introduces performance gains without incurring any computational complexity increase for inference, we advocate it as an alternative to the conventional convolutional layer. We open sourced an implementation of DO-Conv in Tensorflow, PyTorch and GluonCV at https://github.com/yangyanli/DO-Conv.

Keywords:

Kernel

Convolution

Training

Tensors

Color

Computer architecture

Three-dimensional displays

Over-parameterization

convolutional Layer

depthwise convolution

DO-Conv: Depthwise Over-Parameterized Convolutional Layer

Abstract

Journal

Organization