Automatic Classification of Red Blood Cell Morphology Based on Quantitative Phase Imaging

Figure 1

Schematic illustration of the RBC classification process. The process is divided into two stages: training and testing. In the training phase, the model of SSAE trained by the original image block is fine tuned to obtain the DNN. In the testing phase, the BC, DNN, and Softmax classifier are used to classify the cells in the unknown images, where cells A, B, and C are the spherocyte, echinocyte, and discocyte in the prediction results, respectively.