Denoising for single-cell resolution highly multiplexed imaging: Deep learning safeguards cell spatial heterogeneity analysis

High-dimensional single-cell techniques enable exploring the heterogeneous cell populations at single-cell resolution unveiling deep biological insight into various diseases. Single-cell resolution highly multiplexed imaging is one of the cutting-edge techniques for high-dimensional single-cell analysis, preserving spatial context while creating high-dimensional biomarker catalogs of cells. As a part and parcel of highly multiplexed imaging data processing, visualization preprocessing aims to remove noise from original signal, a process achieved through the use of denoising algorithms. While highly multiplexed imaging become more widely adopted, denoising optimizations are critical for advancing future research. In this paper, we summarize the feature of different highly multiplexed imaging techniques and reviews the principles, characteristics, development, and applications of denoising algorithms. Furthermore, we discussed the technical requirements and development prospects, offering assistance for domain researches and technology developers.