Artificial neural network (ANN) modeling and analysis for the prediction of change in the lip curvature following extraction and non-extraction orthodontic treatment

Abstract

Objective: To establish and determine the accuracy of ANN model for the analysis of lip curve change following extraction and non-extraction orthodontic treatment. Methods: Forty adult subjects who required various combinations of premolars extraction and non-extraction for the correction of their malocclusion were chosen. Based on the extraction pattern, all the subjects (n=40) were divided equally into an extraction and a non-extraction group. The effect of extraction and non-extraction treatment on the depth of upper and lower lip curvature was measured on the lateral cephalograms recorded in natural head position. The data obtained from the cephalometric analysis were used to produce a trained ANN model and then the model was analyzed to determine its accuracy in the prediction of upper and lower lip curvature change. Results: The mean change in the depth of upper lip curvature following various combinations of premolars extraction and non-extraction treatment was significantly different (P