Please login first
Cell classification based on artificial intelligence analysis of cell images in microfluidic chip
1 , * 2
1  Chongqing College of Electronic Engineering, Chongqing 401331, China
2  Key Laboratory of Optoelectronic Technology and Systems, Ministry of Education & Key Disciplines Labora-tory of Novel Micro-Nano Devices and System Technology, College of Optoelectronic Engineering, Chongqing University, Chongqing 400044, China
Academic Editor: Sara Tombelli


Biological cells are micro-particles with certain size, shape and state. Most biological cells have complex, irregular appearance and internal structure, organic or chemical composition and biological activity. At present, the classification of biological cells includes traditional optical microscope, fluorescent straining, flow cytometry or PCR, etc. However, these methods suffer several drawbacks such as low sensitivity, expensive and bulky instruments and/or complex processes. With the rapid development of computer algorithms, artificial intelligence (AI) starts to be applied in cell classification, which is simple and does not need any labelling reagent. The main problems of cell classification based on AI are focused on the influence of image background, cell aggregation, small size and lack of enough training samples. We developed a low-cost, multi-classification, label-free and high-precision method for cell classification, which combines microfluidic technology with deep learning algorism together. The recognition of states of red blood cells (RBCs) was selected as the typical example to demonstrate the feasibility of the method. The microfluidic channel is designed to effectively and controllably solve the problem of cell overlap, which has severe negative impact on the identification of cells. The object detection model based on YOLOv5 in the deep learning algorism is optimized and used to recognize multiple RBCs simultaneously in the whole field of view, so as to classify them into six morphological subcategories and count the numbers in each subgroup. The blood quality can be evaluated by calculating the morphology index according to the numbers of cells in subgroups. The validation of the method is verified by evaluating three blood samples stored for 1 week, 3 weeks and 6 weeks, which have distinct morphology index differences. This method has the merit of cell identification in a wide channel, no need for single cell alignment as the image cytometry does and it has potential applications to the classifications of biological cells with different morphologies.

Keywords: Cell classification ; artificial intelligence; object detection model; microfluidic chip