An efficient quantum-classical hybrid algorithm for distorted alphanumeric character identification

TL;DR

This paper introduces a quantum-classical hybrid algorithm that enhances optical character recognition by transforming low-resolution images into high-resolution ones using a variant of Grover's search, demonstrating improved efficiency and confidence.

Contribution

The paper presents a novel hybrid quantum-classical algorithm utilizing fixed point search for OCR, with simulation and complexity analysis showing its advantages over existing methods.

Findings

Achieves higher confidence in character recognition

Demonstrates improved efficiency over classical and quantum algorithms

Utilizes a quantum variant of Grover's search for image enhancement

Abstract

An algorithm for image processing is proposed. The proposed algorithm, which can be viewed as a quantum-classical hybrid algorithm, can transform a low-resolution bitonal image of a character from the set of alphanumeric characters (A-Z, 0-9) into a high-resolution image. The quantum part of the proposed algorithm fruitfully utilizes a variant of Grover's search algorithm, known as the fixed point search algorithm. Further, the quantum part of the algorithm is simulated using CQASM and the advantage of the algorithm is established through the complexity analysis. Additional analysis has also revealed that this scheme for optical character recognition (OCR) leads to high confidence value and generally works in a more efficient manner compared to the existing classical, quantum, and hybrid algorithms for a similar task.