Feature-based Recognition Framework for Super-resolution Images

TL;DR

This paper introduces FGAN, a feature-based recognition network that enhances recognition accuracy on super-resolution images by extracting more beneficial features, outperforming existing models by over 6%.

Contribution

The paper presents a novel feature-based recognition network combined with GANs specifically designed for super-resolution images, improving accuracy over existing models.

Findings

Recognition accuracy increased by more than 6%

Built three datasets using different super-resolution algorithms

FGAN outperforms ReaNet50 and DenseNet121

Abstract

In practical application, the performance of recognition network usually decreases when being applied on super-resolution images. In this paper, we propose a feature-based recognition network combined with GAN (FGAN). Our network improves the recognition accuracy by extracting more features that benefit recognition from SR images. In the experiment, we build three datasets using three different super-resolution algorithm, and our network increases the recognition accuracy by more than 6% comparing with ReaNet50 and DenseNet121.