Deep High-Resolution Representation Learning for Visual Recognition

TL;DR

The paper introduces HRNet, a novel neural network architecture that maintains high-resolution representations throughout the process, leading to improved performance in various vision tasks like pose estimation and segmentation.

Contribution

HRNet is the first to keep high-resolution features in parallel and exchange information across resolutions, enhancing spatial precision and semantic richness.

Findings

HRNet outperforms existing methods in human pose estimation.

HRNet achieves superior results in semantic segmentation.

HRNet serves as a stronger backbone for diverse vision tasks.

Abstract

High-resolution representations are essential for position-sensitive vision problems, such as human pose estimation, semantic segmentation, and object detection. Existing state-of-the-art frameworks first encode the input image as a low-resolution representation through a subnetwork that is formed by connecting high-to-low resolution convolutions \emph{in series} (e.g., ResNet, VGGNet), and then recover the high-resolution representation from the encoded low-resolution representation. Instead, our proposed network, named as High-Resolution Network (HRNet), maintains high-resolution representations through the whole process. There are two key characteristics: (i) Connect the high-to-low resolution convolution streams \emph{in parallel}; (ii) Repeatedly exchange the information across resolutions. The benefit is that the resulting representation is semantically richer and spatially more precise. We show the superiority of the proposed HRNet in a wide range of applications, including human pose estimation, semantic segmentation, and object detection, suggesting that the HRNet is a stronger backbone for computer vision problems. All the codes are available at~{\url{https://github.com/HRNet}}.