Seeing the macro in the micro: a diffusion model-based approach for style transfer in cellular images.

Jiayi CAI; Yong HE; Feng LIU; Byung-Ho KANG; Xuping FENG

Return

Seeing the macro in the micro: a diffusion model-based approach for style transfer in cellular images.

Author: Jiayi CAI ¹ ; Yong HE ¹ ; Feng LIU ² ; Byung-Ho KANG ³ ; Xuping FENG ⁴
Author Information

1. College of Biosystems Engineering and Food Science, Zhejiang University, Hangzhou 310058, China.
2. College of Life Sciences, Nanjing Agricultural University, Nanjing 210095, China.
3. Cellular and Molecular Biology Program, State Key Laboratory of Agrobiotechnology, School of Life Sciences, The Chinese University of Hong Kong, New Territories 999077, Hong Kong, China.
4. College of Biosystems Engineering and Food Science, Zhejiang University, Hangzhou 310058, China. fengxp@zju.edu.cn.
Publication Type:Journal Article
MeSH: Humans; Image Processing, Computer-Assisted/methods*; Cells; Diffusion
From: Journal of Zhejiang University. Science. B 2025;26(6):609-612
CountryChina
Language:English
Abstract: The internal structures of cells as the basic units of life are a major wonder of the microscopic world. Cellular images provide an intriguing window to help explore and understand the composition and function of these structures. Scientific imagery combined with artistic expression can further expand the potential of imaging in educational dissemination and interdisciplinary applications. This study presents an innovative diffusion model-based approach for style transfer in cellular images, combining scientific rigor with artistic expression. By leveraging training-free large-scale pre-trained diffusion models, the proposed method integrates the intricate morphological and textural features of cellular images with diverse artistic styles. Key techniques such as the inversion of denoising diffusion implicit models (DDIMs), adaptive instance normalization (AdaIN), self-attention style injection, and attention temperature scaling ensure the preservation of cellular structures while enhancing visual expressiveness. The results showcase the potential of this strategy for interdisciplinary applications, enriching both the visualization and educational dissemination of cellular imagery through compelling storytelling and aesthetic appeal.