Image Super-resolution via Feature-augmented Random Forest

TL;DR

This paper introduces a feature-augmented random forest method for image super-resolution that enhances features with gradient magnitudes, employs LSH for feature reduction, and uses a weighted ridge regression, outperforming traditional RF and some deep-learning methods.

Contribution

The paper proposes a novel FARF approach combining gradient-augmented features, LSH-based dimensionality reduction, and GWRR, advancing RF-based super-resolution techniques.

Findings

Achieves about 0.3 dB higher PSNR than traditional RF methods.

Outperforms some recent deep-learning super-resolution algorithms.

Effective on multiple benchmark datasets.

Abstract

Recent random-forest (RF)-based image super-resolution approaches inherit some properties from dictionary-learning-based algorithms, but the effectiveness of the properties in RF is overlooked in the literature. In this paper, we present a novel feature-augmented random forest (FARF) for image super-resolution, where the conventional gradient-based features are augmented with gradient magnitudes and different feature recipes are formulated on different stages in an RF. The advantages of our method are that, firstly, the dictionary-learning-based features are enhanced by adding gradient magnitudes, based on the observation that the non-linear gradient magnitude are with highly discriminative property. Secondly, generalized locality-sensitive hashing (LSH) is used to replace principal component analysis (PCA) for feature dimensionality reduction and original high-dimensional features are employed, instead of the compressed ones, for the leaf-nodes' regressors, since regressors can benefit from higher dimensional features. This original-compressed coupled feature sets scheme unifies the unsupervised LSH evaluation on both image super-resolution and content-based image retrieval (CBIR). Finally, we present a generalized weighted ridge regression (GWRR) model for the leaf-nodes' regressors. Experiment results on several public benchmark datasets show that our FARF method can achieve an average gain of about 0.3 dB, compared to traditional RF-based methods. Furthermore, a fine-tuned FARF model can compare to or (in many cases) outperform some recent stateof-the-art deep-learning-based algorithms.