Multi-frame Restoration for High-rate Lissajous Confocal Laser Endomicroscopy

2026-05-01

AI Summary

This study addresses the structured artifacts in high-frame-rate Lissajous confocal laser endomicroscopy (CLE) images, which arise from incomplete pixel sampling due to resonant scanning. To tackle this challenge, the authors present the first paired high-frame-rate Lissajous CLE dataset with spatiotemporal alignment supervision and propose MIRA, a lightweight recurrent network that iteratively fuses multi-frame temporal information through feature reuse and optical flow–based displacement alignment to reconstruct high-quality images. MIRA outperforms both lightweight and computationally intensive state-of-the-art models in image restoration quality while maintaining low computational overhead, making it well-suited for clinical deployment.

Abstract

Lissajous confocal laser endomicroscopy (CLE) is a promising solution for high speed in vivo optical biopsy for handheld scenarios. However, Lissajous scanning traces a resonant trajectory and samples only the visited pixels per frame; at high frame rates, many pixels remain unvisited, creating structured holes. In this work, we introduce the first benchmark for high-rate Lissajous CLE, consisting of low-quality video clips paired with high-quality reference images. The reference images are wide-FOV mosaics obtained by stitching stabilized, slow-scan frames of the same tissue, enabling temporally aligned supervision. Using this dataset, we propose MIRA, a lightweight recurrent framework for Lissajous CLE restoration that iteratively aggregates temporal context through feature reuse and displacement alignment. Our experiments demonstrate that MIRA outperforms both lightweight and high-complexity baselines in restoration quality while maintaining a favorable computational efficiency suitable for clinical deployment.

Problem

• Lissajous scanning
• confocal laser endomicroscopy
• structured holes
• high-frame-rate imaging
• incomplete pixel sampling

Innovation

• Lissajous scanning
• multi-frame restoration
• temporal alignment
• lightweight recurrent network
• confocal laser endomicroscopy