Hashing Image Patches for Zooming

TL;DR

This paper introduces a Bayesian super-resolution algorithm that uses patch-based representation, LLE, and hashing for efficient image zooming, demonstrating improved speed and quality over existing methods.

Contribution

The work's novelty lies in combining hashing with LLE-based Bayesian inference for fast, robust image super-resolution with modular and parallelizable architecture.

Findings

Achieves faster processing times compared to traditional methods.

Provides improved visual quality in super-resolved images.

Demonstrates robustness through comparative experiments.

Abstract

In this paper we present a Bayesian image zooming/super-resolution algorithm based on a patch based representation. We work on a patch based model with overlap and employ a Locally Linear Embedding (LLE) based approach as our data fidelity term in the Bayesian inference. The image prior imposes continuity constraints across the overlapping patches. We apply an error back-projection technique, with an approximate cross bilateral filter. The problem of nearest neighbor search is handled by a variant of the locality sensitive hashing (LSH) scheme. The novelty of our work lies in the speed up achieved by the hashing scheme and the robustness and inherent modularity and parallel structure achieved by the LLE setup. The ill-posedness of the image reconstruction problem is handled by the introduction of regularization priors which encode the knowledge present in vast collections of natural images. We present comparative results for both run-time as well as visual image quality based measurements.