Classification of Germination Images of Pear Pollen Using Random Forest and Convolution Neural Network Models

Verifying pollen germination using microscopic images is a difficult task. It usually time-consuming and may entail reduced accuracy reproducibility. Therefore, in this study, we used random forest (RF) convolutional neural network (CNN) models to perform image classification on raw data corresponding pollens with different rates; the were obtained via flow cytometry. A heat map, which was based RF analysis results, showed that variables significantly influenced decision between NG 60G categories mainly located center top-right regions of 30×30 pixel image. Additionally, variable importance plot among 900 input variables, pixel_316 contributed most toward prediction. Gradient-weighted class activation mapping visualize maps CNN model. The bottom-left region map activated However, not only but also activated. Both classified into high accuracy. considering model does reflect characteristics adjacent more appropriate for classifying various rates distinct classes. Taken together, these results suggest can provide reliable method verifying performance.