Machine Learning Techniques For AFM-Based Imaging of Cells

In recent years, advances in computer vision and machine learning have supported a series of algorithms, and their ability to decipher image contents is impressive. These algorithms are being applied to the study of cell images, and gradually changing the analysis methods of cell image data. Atomic Force Microscopy (AFM) is a powerful tool for imaging and analyzing cell structures. It has been used to study the morphology, structure and mechanical properties of cells. However, it is still a challenging topic to identify cells based on AFM images. This article summarizes the applications of machine learning methods in the cell image analysis by AFM. The research of the cell analysis method begins with the collection of cell data, and it emphasizes to identifying cell types, carcinogenesis or carcinogenesis levels, and its auxiliary work includes the cell segmentation, denoising, super-resolution, generation and simulation. Finally, the research experience in these analysis methods and their guiding significance for further study are discussed.