INTERNATIONAL CONFERENCE-CUM-ROUND TABLE ON TRANSLATIONAL RESEARCH AND INNOVATION IN BEAM TECHNOLOGIES (ICTRIBT-2026)

A Hybrid Quantum–Classical Framework for Image Analysis and Facial Expression Recognition

Not scheduled

20m

Central University of Himachal Pradesh, India

Speakers

Dr Amrita N/A (Patna Women's College)Ms Anshika Raj (Patna Women's College) Anshu Sharma (Patna Women's College)

Description

Abstract
Context: Image analysis and boundary detection are critical in diagnostic and monitoring systems where intensity variations must be resolved under noise and limited resolution. Quantum image pro￾cessing, using superposition and interference principles, offers novel representations and operations for structured analysis, which can complement classical approaches in imaging-intensive experimental setups.
Purpose: This study investigates a hybrid quantum–classical framework for facial image analysis and facial expression recognition, explicitly motivated by applications where accurate detection of spatial patterns is crucial for experimental diagnostics and monitoring.
Methods: Facial images are enhanced using histogram equalization and adaptive thresholding to maximize contrast and improve separability prior to quantum-inspired processing. The enhanced images are then encoded using a gridding-based representation analogous to block-based quantum image models, enabling localized subsystem analysis. Facial expressions are characterized through graph-based observables derived from facial landmark points, and classification is performed using classical and quantum-inspired classifiers.
Findings: Thresholding-based preprocessing improves image quality by enhancing contrast and segmentation boundaries, while gridding-based segmentation enables localized analysis and stable feature extraction. Experiments conducted on a subset of the FFHQ dataset achieve approximately 90% accuracy in small-scale evaluations. The workflow demonstrates potential for reliable pattern recognition in image-intensive diagnostic environments.
Significance: The results demonstrate that hybrid quantum–classical frameworks can effectively bridge classical image processing and quantum-inspired algorithms, providing scalable and robust tools for monitoring, diagnostics, and structured analysis in imaging-intensive experimental systems.
Keywords: Quantum image processing; hybrid quantum–classical frameworks; quantum-inspired mod￾els; image gridding; experimental diagnostics; pattern recognition
References
[1] Wang X, Zhao Y, Liu H 2022 IEEE Access 10 42167–42188
[2] Venegas-Andraca S E 2023 Quantum Information Processing 22(3) 85
[3] Mogos D, Radu A 2024 Quantum Information Processing 23(1) 15
[4] Geng L, Li Q, Sun Y 2022 Information Sciences 608 1423–1436
[5] Liu R, Zhang K 2022 Quantum Information Processing 21(4) 112