Real Image Super Resolution Via Heterogeneous Model Ensemble using GP-NAS

TL;DR

This paper introduces a Gaussian process-based neural architecture search to develop diverse models for real image super-resolution, achieving top results in the AIM 2020 challenge.

Contribution

It proposes a novel GP-NAS framework to automatically design heterogeneous models for real image SR, outperforming existing methods.

Findings

Achieved first place in all three tracks of AIM 2020 Real Image Super-Resolution Challenge.

Developed a suite of diverse models through GP-NAS for effective ensemble performance.

Demonstrated superior performance on real-world datasets compared to prior approaches.

Abstract

With advancement in deep neural network (DNN), recent state-of-the-art (SOTA) image superresolution (SR) methods have achieved impressive performance using deep residual network with dense skip connections. While these models perform well on benchmark dataset where low-resolution (LR) images are constructed from high-resolution (HR) references with known blur kernel, real image SR is more challenging when both images in the LR-HR pair are collected from real cameras. Based on existing dense residual networks, a Gaussian process based neural architecture search (GP-NAS) scheme is utilized to find candidate network architectures using a large search space by varying the number of dense residual blocks, the block size and the number of features. A suite of heterogeneous models with diverse network structure and hyperparameter are selected for model-ensemble to achieve outstanding performance in real image SR. The proposed method won the first place in all three tracks of the AIM 2020 Real Image Super-Resolution Challenge.