Research on Recognition and Classification System of Cerebrospinal Fluid Cells Based on Small Samples

This article introduces the method of identifying and classifying small samples of cerebrospinal fluid cells. There are many uncertain influencing factors in the extraction, preparation and sampling of cerebrospinal fluid cells, resulting in some irrelevant factors in the sample other than the target cells that affect the subsequent identification and classification. In this paper, three main types of cells are identified in human cerebrospinal fluid cells, lymphocytes, monocytes and centrioles. First, preprocess the pictures for the background impurities caused by lens contamination or improper operation in the collected cell samples; secondly, due to the characteristics of the samples with complex backgrounds that there are not many high-quality pictures, artificial neural networks are used. The training of the model on a small sample of cell image sets has achieved good results.