AFM-Based Research On Recognition Of Biological Features Of Cells

Atomic Force Microscopy (AFM) has been applied for building nano-devices for almost a decade. It also offers the prospect for researching and measuring biological objects at nanometer scale. Utilizing AFM’s scanning topographical images, features were extracted and recognized on a great number of different cells in this paper. According to topographical features of different cells types, wavelet transform or Laplace transform were selected to combine with histogram equalization and noise filtering methods to enhance and preprocess the topographical images, and then after morphology operation and threshold segmentation, the optimum cells sample could been separated. More importantly, by considering the cell image with its particular configuration, the3-D features (Surface Roughness of Cells, Surface Area of Cells and Volume of Cells) and2-D projection features (Perimeter, Area and Degree of Circularity) were extracted from the sample. Finally, those features were used as criteria to evaluate the cell’s performance.There were more than180images used to test and evaluate those methods, which presented in this paper, including46topographical images of different plant cells,64topographical images of normal spleen cells and70topographical images of abnormal spleen cells, and returned almost91.2%accuracy even in different complex environment. It was demonstrated that those methods can be used to judge cells with great accuracy. These biomedical investigations with features recognition base on AFM can offer more refined theory support in diagnosis and treatment of diseases and facilitated the biological features extraction technique base on AFM in real time.