Partial Convolution based Padding

TL;DR

This paper introduces a novel padding method for convolutional neural networks called partial convolution based padding, which improves accuracy by re-weighting border regions during convolution.

Contribution

It proposes a simple padding scheme that treats padded regions as holes and adjusts convolution results accordingly, outperforming standard zero padding.

Findings

Consistently improves accuracy over zero padding on ImageNet classification.

Effective for various deep network models and tasks like semantic segmentation.

Enhances border handling in convolutional operations.

Abstract

In this paper, we present a simple yet effective padding scheme that can be used as a drop-in module for existing convolutional neural networks. We call it partial convolution based padding, with the intuition that the padded region can be treated as holes and the original input as non-holes. Specifically, during the convolution operation, the convolution results are re-weighted near image borders based on the ratios between the padded area and the convolution sliding window area. Extensive experiments with various deep network models on ImageNet classification and semantic segmentation demonstrate that the proposed padding scheme consistently outperforms standard zero padding with better accuracy.