Noise-Embedded Image Processing Based on Quantum Data Encodings

Advancements in quantum information have significantly impacted the field of image processing, although challenges remain. Especially in the edge detection and image encoding area, distorted feature and noises would affect the further classification or super resolution tasks. In our work, we conduct researches on two stages to both evaluate the potential of Quantum-based Convolutional Structure in extracting distorted feature and further explore the effects of quantum noise channels on quantum image encodings. In the first stage, we propose a method to extract distorted edge features by applying shallow layers in quantum convolutional neural networks (QCNN). By combining the advantages of quantum computing and the layered structure of convolutional neural networks (CNN), this approach addresses the problem and compares the extracted distorted pattern with the reference pattern, achieving a best matching ratio of 99.71% in images interfered with impulse noise. Furthermore, we also compare the performance with the classical method and other quantum algorithms, our method gains a 97.19% ratio. In the second stage, we focus on quantum noises interfere with the circuits-based quantum image encoding that store classical image to quantum machine. We explore the potential effects of quantum noise models on varied encodings, combining both quantum evaluation methodology such as circuit depth, qubit number, fidelity comparison and depth growth, and classical image evaluation method such as mean squared error (MSE), structural similarity index measure (SSIM) and peak signal-noise ratio (PSNR) to find the pattern of noise behaviors affecting encodings. We also evaluate the behavior in applying noise to specific gate sets and measurement to distinguish how noise would affect the circuit. In addition, two datasets are evaluated to compare the effects on images with different levels of complexity.