Figure 5.

Classifier performance metrics.

(A) shows the performance of our classifier. We compare performance at different stages of our classifier, starting with the XgBoost classifier, adding an HMM model, increasing features to include 7 Hu moments, and integrating SPINDLE annotations to improve epoch quality. We can see the overall accuracy improves by adding each of these steps. (B) The top 20 most important features for the classifier. (C) Confusion matrix obtained from 10-fold cross validation. (D) Precision-recall table.