Improving the Classification Accuracy in Label-Free Flow Cytometry Using Event-Based Vision and Simple Logistic Regression