Studying Very Low Resolution Recognition Using Deep Networks

TL;DR

This paper introduces a deep learning approach called Robust Partially Coupled Networks to improve recognition accuracy in Very Low Resolution Recognition (VLRR) tasks, addressing challenges of small ROI sizes and domain mismatch.

Contribution

It proposes a novel deep learning framework that combines super resolution, domain adaptation, and robust regression techniques specifically for VLRR problems.

Findings

Achieved high accuracy on face identification, digit, and font recognition tasks.

Demonstrated robustness to outliers and domain mismatch.

Validated effectiveness through extensive experiments.

Abstract

Visual recognition research often assumes a sufficient resolution of the region of interest (ROI). That is usually violated in practice, inspiring us to explore the Very Low Resolution Recognition (VLRR) problem. Typically, the ROI in a VLRR problem can be smaller than $16 \times 16$ pixels, and is challenging to be recognized even by human experts. We attempt to solve the VLRR problem using deep learning methods. Taking advantage of techniques primarily in super resolution, domain adaptation and robust regression, we formulate a dedicated deep learning method and demonstrate how these techniques are incorporated step by step. Any extra complexity, when introduced, is fully justified by both analysis and simulation results. The resulting \textit{Robust Partially Coupled Networks} achieves feature enhancement and recognition simultaneously. It allows for both the flexibility to combat the LR-HR domain mismatch, and the robustness to outliers. Finally, the effectiveness of the proposed models is evaluated on three different VLRR tasks, including face identification, digit recognition and font recognition, all of which obtain very impressive performances.