Talk: Quantum Edge Detection for Industrial X-ray Images on NISQ Hardware

Organizer:
Lehrstuhl für Experimentalphysik (Eichler)

Speaker:
Anastasia Papadaki

Title:
Quantum Edge Detection for Industrial X-ray Images on NISQ Hardware

Abstract:
Edge detection is a key building block in industrial X-ray computed tomography (XCT), supporting segmentation, defect localization, and feature extraction under noisy acquisition conditions. In this work, we present a hardware-oriented quantum edge-detection framework that translates classical Haar wavelet and Sobel operators into shallow quantum circuits suitable for NISQ devices. The method operates on overlapping 2×2 image patches encoded via angle-based RY rotations and processed using a compact bidirectional quantum Haar transform. Directional sensitivity is introduced through small fixed phase shifts (π/8) applied to selected Haar detail subbands, emulating Sobel-like gradient extraction. The resulting four-qubit circuits have a logical depth below 20 and are evaluated on real industrial XCT radiographs. Across multiple simulators and IBM quantum hardware, the quantum edge maps show strong agreement with classical references, demonstrating that meaningful edge information can be extracted within current hardware constraints.