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

• Author Details:   
• Smruti Bhusan Nanda
• Anmol S Kalha
• Ashok Kumar Jena
• Virag Bhatia
• Sumita Mishra

---

View PDF

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<0.05). The predicted values of upper and lower lip curvature change by ANN model were very close to the actual regression analysis values. However, the mean error in predicting the change in the upper and lower lip curvature by ANN model analysis was only 29.6% and 7% respectively which was much less as compared to the routine regression analysis.
Conclusions: The premolars extraction and non-extraction orthodontic treatment had significant effect on the depth of upper lip curve, and the mean error in predicting the change in lip curvature with ANN analysis was much less as compared to computer based statistical analysis.

Key words: Lip curvature changes, Extraction and non-extraction treatment, Artificial neural network analysis.