1. [Publications](/index.php/publications)
2. Pixel-Adaptive Convolutional Neural Networks
 
 # Pixel-Adaptive Convolutional Neural Networks

  ![](/sites/default/files/styles/wide/public/publications/pac.png?itok=Gjxafk4h)

 We propose a pixel-adaptive convolution (PAC) operation, a simple yet effective modification of standard convolutions, in which the filter weights are multiplied with a spatially-varying kernel that depends on learnable, local pixel features. PAC is a generalization of several popular filtering techniques and thus can be used for a wide range of use cases. Specifically, we demonstrate state-of-the-art performance when PAC is used for deep joint image upsampling. PAC also offers an effective alternative to fully-connected CRF (Full-CRF), called PAC-CRF, which performs competitively, while being considerably faster. In addition, we also demonstrate that PAC can be used as a drop-in replacement for convolution layers in pre-trained networks, resulting in consistent performance improvements.



 ## Authors



Hang Su (University of Massachusetts, Amherst)

Varun Jampani (NVIDIA)

Deqing Sun (NVIDIA)

Orazio Gallo (NVIDIA)

Erik-Learned Miller (University of Massachusetts, Amherst)

[Jan Kautz](/index.php/person/jan-kautz)

 

 

 ## Publication Date



Sunday, June 16, 2019

 

 ## Published in



[Computer Vision and Pattern Recognition (CVPR), 2019](http://cvpr2019.thecvf.com)

 

 ## Research Area



[Computer Vision](/index.php/research-area/computer-vision)

[Artificial Intelligence and Machine Learning ](/index.php/research-area/machine-learning-artificial-intelligence)

 

 

 ## External Links



[PDF](https://arxiv.org/pdf/1904.05373.pdf)

[Project Page](https://suhangpro.github.io/pac/index.html)

[Code](https://github.com/NVlabs/pacnet)

[Video Intro](https://youtu.be/gsQZbHuR64o)

 

 

 ## Copyright



This material is posted here with permission of the IEEE. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to <pubs-permissions@ieee.org>.