Partial Convolution based Padding

By Guilin Liu, Kevin J. Shih, Ting-Chun Wang, Fitsum A. Reda, Karan Sapra, Zhiding Yu, Andrew Tao, Bryan Catanzaro. 2018

101 citations recorded by the source corpus.

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.

Subfields: Computer Vision and Pattern Recognition, Model Architectures

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Model Architectures

Citations in the corpus

Global Pooling, More than Meets the Eye: Position Information is Encoded Channel-Wise in CNNs (2021, background): “The ability of zero padding to inject position information has been an area of focus in multiple follow-up works [18, 1, 16, 23, 24].”
Mind the Pad - CNNs can Develop Blind Spots (2020, methodology): “Partial Convolution [20] has been proposed as a padding method that treats pixels outside the original image as missing values and rescales the computed convolutions accordingly [21].”
On Translation Invariance in CNNs: Convolutional Layers Can Exploit Absolute Spatial Location (2020, background): “Boundary handling in CNNs focuses on minimizing boundary effects by learning separate filters at the boundary [44], treating out of boundary pixels as missing values [62], circular convolutions for wrap-around input data such as 360◦ degree images [85] and minimizing distortions in 360◦ degree video”